Aragora API Reference

Last Updated: 2026-02-03 (v2.6.3 alignment with repo versions)

This document describes the HTTP and WebSocket APIs for Aragora's control plane for multi-agent vetted decisionmaking across organizational knowledge and channels.

SDK Coverage

SDK	Version	Methods	Coverage
TypeScript (@aragora/sdk)	2.6.3	380 async	Full API (79 namespaces)
Python (aragora)	2.6.3	220 async + 420 sync	Full API

Versions reflect the current repo tags (see pyproject.toml and aragora/__version__.py). If versions drift, run python scripts/check_version_alignment.py.

Both SDKs provide complete coverage of all API endpoints including:

• Debates, Agents, Memory, Knowledge
• Gauntlet, Verification, Workflows
• Control Plane, RBAC, Tenancy
• Authentication, Billing, Audit
• Backups, Expenses, RLM, Unified Inbox, Feedback

Related Documentation

Document	Purpose	When to Use
API_REFERENCE.md (this)	Complete endpoint catalog	Primary reference for all endpoints
API_ENDPOINTS.md	Auto-generated endpoint list	Cross-check handler coverage
API_EXAMPLES.md	Runnable code examples	Learning API patterns
API_RATE_LIMITS.md	Rate limiting details	Understanding throttling
API_VERSIONING.md	Version strategy	API migration planning
API_STABILITY.md	Stability guarantees	Production decisions
WEBSOCKET_EVENTS.md	WebSocket event types	Real-time integration

Endpoint Usage Status

Endpoint counts vary by deployment and enabled handlers. To audit the current surface area, run:

python scripts/generate_api_docs.py --format json

OpenAPI Specification

The generated OpenAPI spec lives in docs/api/openapi.json and docs/api/openapi.yaml (JSON-formatted for compatibility). Regenerate them with:

python scripts/export_openapi.py --output-dir docs/api

The canonical spec is produced by aragora/server/openapi and the endpoint definitions under aragora/server/openapi/endpoints/. If you add or change endpoints, update the OpenAPI endpoint definitions and re-export the docs.

Each OpenAPI operation includes x-aragora-stability to indicate whether it is stable, beta, experimental, internal, or deprecated. See docs/API_STABILITY.md for the promotion workflow.

Auth Signup & SSO API

Self-service signup and SSO/OIDC endpoints live under /api/v1/auth. Signup flows are in-memory by default; use a database-backed store for production.

Method	Endpoint	Description
POST	/api/v1/auth/signup	Register new user
POST	/api/v1/auth/verify-email	Verify email address
POST	/api/v1/auth/resend-verification	Resend verification email
POST	/api/v1/auth/setup-organization	Create org after signup
POST	/api/v1/auth/invite	Invite team member
POST	/api/v1/auth/accept-invite	Accept invitation
GET	/api/v1/auth/check-invite	Check invitation validity
GET	/api/v1/auth/sso/login	Get SSO authorization URL
GET	/api/v1/auth/sso/callback	Handle OAuth/OIDC callback
POST	/api/v1/auth/sso/refresh	Refresh access token
POST	/api/v1/auth/sso/logout	Logout from SSO
GET	/api/v1/auth/sso/providers	List available providers
GET	/api/v1/auth/sso/config	Get provider configuration

SCIM 2.0 Provisioning API

SCIM 2.0 (RFC 7643/7644) endpoints for automated user and group provisioning from identity providers (Okta, Azure AD, OneLogin, etc.). Endpoints are mounted at /scim/v2/ and use Bearer token authentication.

Authentication: Set SCIM_BEARER_TOKEN environment variable. All requests must include Authorization: Bearer <token> header.

User Endpoints

Method	Endpoint	Description
GET	/scim/v2/Users	List users with filtering and pagination
POST	/scim/v2/Users	Create a new user
GET	/scim/v2/Users/\{id\}	Get user by ID
PUT	/scim/v2/Users/\{id\}	Replace user (full update)
PATCH	/scim/v2/Users/\{id\}	Partial update user
DELETE	/scim/v2/Users/\{id\}	Delete user (soft delete by default)

Group Endpoints

Method	Endpoint	Description
GET	/scim/v2/Groups	List groups with filtering and pagination
POST	/scim/v2/Groups	Create a new group
GET	/scim/v2/Groups/\{id\}	Get group by ID
PUT	/scim/v2/Groups/\{id\}	Replace group (full update)
PATCH	/scim/v2/Groups/\{id\}	Partial update group (add/remove members)
DELETE	/scim/v2/Groups/\{id\}	Delete group

Query Parameters

Parameter	Type	Description	Default
startIndex	int	1-based pagination offset	1
count	int	Page size (max 1000)	100
filter	string	SCIM filter expression	—

Supported Filter Operators

eq, ne, co, sw, ew, pr, gt, ge, lt, le, and, or

Example: filter=userName eq "john@example.com"

Content Type

All responses use application/scim+json media type per RFC 7644.

Dashboard API

Dashboard endpoints return overview stats, activity, and quick actions. Responses are cached in-memory for 30 seconds by default.

Method	Endpoint	Description
GET	/api/v1/dashboard	Get dashboard overview
GET	/api/v1/dashboard/stats	Get detailed stats
GET	/api/v1/dashboard/activity	Get recent activity
GET	/api/v1/dashboard/inbox-summary	Get inbox summary
GET	/api/v1/dashboard/quick-actions	Get available quick actions
POST	/api/v1/dashboard/quick-actions/\{action\}	Execute quick action

Deliberations API

Active vetted decisionmaking sessions and stats.

Method	Endpoint	Description
GET	/api/v1/deliberations/active	List active deliberations
GET	/api/v1/deliberations/stats	Deliberation statistics
GET	/api/v1/deliberations/\{id\}	Get deliberation details
GET	/api/v1/deliberations/stream	WebSocket stream of deliberations

Code Review API

Code review endpoints run multi-agent reviews for snippets, diffs, and PRs. Results are stored in-memory by default.

Method	Endpoint	Description
POST	/api/v1/code-review/review	Review code snippet
POST	/api/v1/code-review/diff	Review diff/patch
POST	/api/v1/code-review/pr	Review GitHub PR
GET	/api/v1/code-review/results/\{id\}	Get review result
GET	/api/v1/code-review/history	Review history
POST	/api/v1/code-review/security-scan	Quick security scan

Devices API

Device registration and push notification endpoints.

Method	Endpoint	Description
POST	/api/v1/devices/register	Register a device for push notifications
GET	/api/v1/devices/health	Device connector health
GET	/api/v1/devices/user/\{user_id\}	List user devices
POST	/api/v1/devices/user/\{user_id\}/notify	Notify all user devices
GET	/api/v1/devices/\{device_id\}	Get device info
DELETE	/api/v1/devices/\{device_id\}	Unregister device
POST	/api/v1/devices/\{device_id\}/notify	Notify a device
POST	/api/v1/devices/alexa/webhook	Alexa skill webhook
POST	/api/v1/devices/google/webhook	Google Home webhook

A2A (Agent-to-Agent) API

Agent discovery and task execution endpoints.

Method	Endpoint	Description
GET	/api/v1/a2a/.well-known/agent.json	Agent discovery card
GET	/api/v1/a2a/openapi.json	A2A OpenAPI spec
GET	/api/v1/a2a/agents	List agents
GET	/api/v1/a2a/agents/\{name\}	Get agent details
POST	/api/v1/a2a/tasks	Submit task
GET	/api/v1/a2a/tasks/\{id\}	Get task status
POST	/api/v1/a2a/tasks/\{id\}/stream	Stream task output

Metrics API

Operational and Prometheus metrics endpoints.

Method	Endpoint	Description
GET	/api/metrics	Operational metrics
GET	/api/metrics/health	Health metrics
GET	/api/metrics/cache	Cache metrics
GET	/api/metrics/system	System metrics
GET	/api/metrics/verification	Verification metrics
GET	/api/metrics/background	Background task metrics
GET	/api/metrics/debate	Debate metrics
GET	/metrics	Prometheus export

Plugins API

Plugin management and marketplace endpoints.

Method	Endpoint	Description
GET	/api/v1/plugins	List plugins
GET	/api/v1/plugins/installed	List installed plugins
GET	/api/v1/plugins/marketplace	Marketplace listings
GET	/api/v1/plugins/submissions	List submissions
POST	/api/v1/plugins/submit	Submit plugin
GET	/api/v1/plugins/\{name\}	Plugin details
POST	/api/v1/plugins/\{name\}/install	Install plugin
DELETE	/api/v1/plugins/\{name\}/install	Uninstall plugin
POST	/api/v1/plugins/\{name\}/run	Run plugin action

Cross-Pollination API

Knowledge sharing between debates and Knowledge Mound.

Method	Endpoint	Description
GET	/api/v1/cross-pollination/stats	Cross-pollination stats
GET	/api/v1/cross-pollination/metrics	Cross-pollination metrics
GET	/api/v1/cross-pollination/subscribers	List subscribers
POST	/api/v1/cross-pollination/subscribe	Subscribe a debate
DELETE	/api/v1/cross-pollination/subscribers/\{debate_id\}	Unsubscribe a debate
GET	/api/v1/cross-pollination/bridge	Bridge configuration
PUT	/api/v1/cross-pollination/bridge	Update bridge config
GET	/api/v1/cross-pollination/km	Knowledge Mound status
GET	/api/v1/cross-pollination/km/staleness-check	Knowledge staleness check
GET	/api/v1/laboratory/cross-pollinations/suggest	Suggest cross-pollination

Platform APIs (Advertising, Analytics, CRM, Ecommerce, Support)

These endpoints unify operational data across platforms. They are backed by SecureHandler and respect RBAC permissions.

Advertising

Method	Endpoint	Description
GET	/api/v1/advertising/platforms	List connected platforms
POST	/api/v1/advertising/connect	Connect a platform
DELETE	/api/v1/advertising/\{platform\}	Disconnect platform
GET	/api/v1/advertising/campaigns	Cross-platform campaigns
POST	/api/v1/advertising/\{platform\}/campaigns	Create campaign
GET	/api/v1/advertising/performance	Cross-platform performance
POST	/api/v1/advertising/analyze	Performance analysis

Analytics

Method	Endpoint	Description
GET	/api/v1/analytics/platforms	List connected platforms
POST	/api/v1/analytics/connect	Connect a platform
GET	/api/v1/analytics/dashboards	Cross-platform dashboards
POST	/api/v1/analytics/query	Execute unified query
GET	/api/v1/analytics/reports	Available reports
POST	/api/v1/analytics/reports/generate	Generate report
GET	/api/v1/analytics/metrics	Metrics overview
GET	/api/v1/analytics/realtime	Real-time metrics

Endpoint Analytics

Method	Endpoint	Description
GET	/api/analytics/endpoints	Aggregated endpoint metrics
GET	/api/analytics/endpoints/slowest	Top N slowest endpoints
GET	/api/analytics/endpoints/errors	Top N endpoints by error rate
GET	/api/analytics/endpoints/\{endpoint\}/performance	Endpoint performance detail
GET	/api/analytics/endpoints/health	Overall API health summary

CRM

Method	Endpoint	Description
GET	/api/v1/crm/platforms	List connected platforms
POST	/api/v1/crm/connect	Connect a platform
GET	/api/v1/crm/contacts	Cross-platform contacts
POST	/api/v1/crm/\{platform\}/contacts	Create contact
GET	/api/v1/crm/companies	Companies
GET	/api/v1/crm/deals	Deals/opportunities
GET	/api/v1/crm/pipeline	Sales pipeline
POST	/api/v1/crm/sync-lead	Sync lead
POST	/api/v1/crm/enrich	Enrich contact data

Ecommerce

Method	Endpoint	Description
GET	/api/v1/ecommerce/platforms	List connected platforms
POST	/api/v1/ecommerce/connect	Connect a platform
GET	/api/v1/ecommerce/orders	Cross-platform orders
GET	/api/v1/ecommerce/products	Products
GET	/api/v1/ecommerce/inventory	Inventory levels
POST	/api/v1/ecommerce/sync-inventory	Sync inventory
GET	/api/v1/ecommerce/fulfillment	Fulfillment status
POST	/api/v1/ecommerce/ship	Create shipment

Support

Method	Endpoint	Description
GET	/api/v1/support/platforms	List connected platforms
POST	/api/v1/support/connect	Connect a platform
GET	/api/v1/support/tickets	Cross-platform tickets
POST	/api/v1/support/triage	Ticket triage
POST	/api/v1/support/auto-respond	Response suggestions

Codebase Analysis API

Codebase security and metrics endpoints live under /api/v1/codebase and /api/v1/cve. For full examples and response shapes, see docs/CODEBASE_ANALYSIS.md.

Method	Endpoint	Description
POST	/api/v1/codebase/\{repo\}/scan	Run dependency vulnerability scan
GET	/api/v1/codebase/\{repo\}/scan/latest	Latest scan result
GET	/api/v1/codebase/\{repo\}/scan/\{scan_id\}	Scan by ID
GET	/api/v1/codebase/\{repo\}/scans	List scan history
GET	/api/v1/codebase/\{repo\}/vulnerabilities	Vulnerabilities list
GET	/api/v1/codebase/package/\{ecosystem\}/\{package\}/vulnerabilities	Package advisories
GET	/api/v1/cve/\{cve_id\}	CVE details
POST	/api/v1/codebase/\{repo\}/scan/sast	Trigger SAST scan
GET	/api/v1/codebase/\{repo\}/scan/sast/\{scan_id\}	SAST scan status
GET	/api/v1/codebase/\{repo\}/sast/findings	SAST findings
GET	/api/v1/codebase/\{repo\}/sast/owasp-summary	OWASP summary
POST	/api/v1/codebase/\{repo\}/metrics/analyze	Run metrics analysis
GET	/api/v1/codebase/\{repo\}/metrics	Latest metrics report
GET	/api/v1/codebase/\{repo\}/metrics/\{analysis_id\}	Metrics report by ID
GET	/api/v1/codebase/\{repo\}/metrics/history	Metrics history
GET	/api/v1/codebase/\{repo\}/hotspots	Complexity hotspots
GET	/api/v1/codebase/\{repo\}/duplicates	Code duplication summary
GET	/api/v1/codebase/\{repo\}/metrics/file/\{file_path\}	File-level metrics

Additional codebase analysis endpoints:

Method	Endpoint	Description
POST	/api/v1/codebase/analyze-dependencies	Analyze dependencies
POST	/api/v1/codebase/scan-vulnerabilities	Scan repo for CVEs
POST	/api/v1/codebase/check-licenses	License compatibility
POST	/api/v1/codebase/sbom	Generate SBOM
POST	/api/v1/codebase/clear-cache	Clear dependency analysis cache
POST	/api/v1/codebase/\{repo\}/scan/secrets	Trigger secrets scan
GET	/api/v1/codebase/\{repo\}/scan/secrets/latest	Latest secrets scan
GET	/api/v1/codebase/\{repo\}/scan/secrets/\{scan_id\}	Secrets scan by ID
GET	/api/v1/codebase/\{repo\}/secrets	Secrets list
GET	/api/v1/codebase/\{repo\}/scans/secrets	Secrets scan history

Code Intelligence API

Code intelligence endpoints support AST analysis, call graphs, and audit runs. Canonical endpoints live under /api/v1/codebase with /api/codebase aliases for UI-driven flows.

Method	Endpoint	Description
POST	/api/v1/codebase/\{repo\}/analyze	Analyze codebase structure
GET	/api/v1/codebase/\{repo\}/symbols	List symbols
GET	/api/v1/codebase/\{repo\}/callgraph	Fetch call graph
GET	/api/v1/codebase/\{repo\}/deadcode	Find dead code
POST	/api/v1/codebase/\{repo\}/impact	Impact analysis
POST	/api/v1/codebase/\{repo\}/understand	Answer questions about code
POST	/api/v1/codebase/\{repo\}/audit	Run comprehensive audit
GET	/api/v1/codebase/\{repo\}/audit/\{audit_id\}	Audit status/result

Quick Scan API

Quick scan endpoints power the one-click security scan UI.

Method	Endpoint	Description
POST	/api/codebase/quick-scan	Run quick security scan
GET	/api/codebase/quick-scan/\{scan_id\}	Get quick scan result
GET	/api/codebase/quick-scans	List quick scans

GitHub PR Review API

GitHub pull request review endpoints live under /api/v1/github/pr. For full details, see docs/GITHUB_PR_REVIEW.md. Register PRReviewHandler in the server handler registry to expose these routes in the unified server.

Method	Endpoint	Description
POST	/api/v1/github/pr/review	Trigger a PR review
GET	/api/v1/github/pr/\{pr_number\}	Get PR details
GET	/api/v1/github/pr/review/\{review_id\}	Get review status/result
GET	/api/v1/github/pr/\{pr_number\}/reviews	List reviews for a PR
POST	/api/v1/github/pr/\{pr_number\}/review	Submit review to GitHub

GitHub Audit Bridge API

Audit-to-GitHub endpoints live under /api/v1/github/audit.

Method	Endpoint	Description
POST	/api/v1/github/audit/issues	Create issues from findings
POST	/api/v1/github/audit/issues/bulk	Bulk create issues
POST	/api/v1/github/audit/pr	Create PR with fixes
GET	/api/v1/github/audit/sync/\{session_id\}	Get sync status
POST	/api/v1/github/audit/sync/\{session_id\}	Sync session to GitHub

Audit Sessions API

Audit session endpoints manage multi-agent document audits. The UI uses /api/audit/sessions; versioned endpoints live under /api/v1/audit/sessions.

Method	Endpoint	Description
POST	/api/v1/audit/sessions	Create audit session
GET	/api/v1/audit/sessions	List audit sessions
GET	/api/v1/audit/sessions/\{session_id\}	Get session details
DELETE	/api/v1/audit/sessions/\{session_id\}	Delete session
POST	/api/v1/audit/sessions/\{session_id\}/start	Start audit
POST	/api/v1/audit/sessions/\{session_id\}/pause	Pause audit
POST	/api/v1/audit/sessions/\{session_id\}/resume	Resume audit
POST	/api/v1/audit/sessions/\{session_id\}/cancel	Cancel audit
GET	/api/v1/audit/sessions/\{session_id\}/findings	List findings
GET	/api/v1/audit/sessions/\{session_id\}/events	Stream SSE events
POST	/api/v1/audit/sessions/\{session_id\}/intervene	Human intervention
GET	/api/v1/audit/sessions/\{session_id\}/report	Export report

Security Debate API

Trigger multi-agent debates on security vulnerability findings. Debates use the Arena with security-focused agents to analyze vulnerabilities and recommend remediation strategies.

Authentication: Required. Permissions: audit:write (POST), audit:read (GET). Rate limit: 10 requests per minute per user.

Method	Endpoint	Description
POST	/api/v1/audit/security/debate	Trigger security debate on findings
GET	/api/v1/audit/security/debate/\{id\}	Get security debate status

POST /api/v1/audit/security/debate

Trigger a multi-agent security debate on vulnerability findings.

Request body:

{
    "findings": [
        {
            "id": "optional-uuid",
            "finding_type": "vulnerability",
            "severity": "critical",
            "title": "SQL Injection in user handler",
            "description": "Unsanitized user input passed to SQL query",
            "file_path": "aragora/server/handlers/users.py",
            "line_number": 42,
            "cve_id": "CVE-2024-1234",
            "package_name": "optional-package",
            "package_version": "1.2.3",
            "recommendation": "Use parameterized queries",
            "metadata": {}
        }
    ],
    "repository": "repo-name",
    "confidence_threshold": 0.7,
    "timeout_seconds": 300
}

Field	Type	Required	Description
findings	array	Yes	List of security findings to debate
repository	string	No	Repository name (default: "unknown")
confidence_threshold	number	No	Debate confidence threshold, 0.1-1.0 (default: 0.7)
timeout_seconds	integer	No	Debate timeout, 30-600 (default: 300)

Response (200):

{
    "debate_id": "uuid",
    "status": "completed",
    "consensus_reached": true,
    "confidence": 0.85,
    "final_answer": "Remediation recommendations...",
    "rounds_used": 3,
    "duration_ms": 12500,
    "findings_analyzed": 2,
    "votes": {
        "SecurityAnalyst": "remediate",
        "CodeReviewer": "remediate"
    }
}

GET /api/v1/audit/security/debate/{id}

Get the status of a previously triggered security debate. Currently debates are synchronous, so this endpoint returns not_found for any ID (placeholder for future async support).

Response (200):

{
    "debate_id": "the-id",
    "status": "not_found",
    "message": "Debate results are not persisted. Use POST to trigger a new debate."
}

Shared Inbox API

Shared inbox endpoints live under /api/v1/inbox. For workflow details, see docs/SHARED_INBOX.md.

Method	Endpoint	Description
POST	/api/v1/inbox/shared	Create shared inbox
GET	/api/v1/inbox/shared	List shared inboxes
GET	/api/v1/inbox/shared/\{id\}	Get inbox details
GET	/api/v1/inbox/shared/\{id\}/messages	List inbox messages
POST	/api/v1/inbox/shared/\{id\}/messages/\{msg_id\}/assign	Assign message
POST	/api/v1/inbox/shared/\{id\}/messages/\{msg_id\}/status	Update status
POST	/api/v1/inbox/shared/\{id\}/messages/\{msg_id\}/tag	Add tag
POST	/api/v1/inbox/routing/rules	Create routing rule
GET	/api/v1/inbox/routing/rules	List routing rules
PATCH	/api/v1/inbox/routing/rules/\{id\}	Update routing rule
DELETE	/api/v1/inbox/routing/rules/\{id\}	Delete routing rule
POST	/api/v1/inbox/routing/rules/\{id\}/test	Test routing rule

Inbox Command Center API

Command center endpoints power the inbox UI quick actions and daily digest.

Method	Endpoint	Description
GET	/api/inbox/command	Fetch prioritized inbox
POST	/api/inbox/actions	Execute quick action
POST	/api/inbox/bulk-actions	Execute bulk action
POST	/api/inbox/reprioritize	Trigger AI re-prioritization
GET	/api/inbox/sender-profile	Sender profile
GET	/api/inbox/daily-digest	Daily digest
GET	/api/v1/inbox/daily-digest	Daily digest (v1 alias)
GET	/api/email/daily-digest	Daily digest (email alias)

Message Bindings API

Bindings map messages to routing targets (provider/account/pattern). All routes support /api/v1 via version stripping.

Method	Endpoint	Description
GET	/api/bindings	List all bindings
GET	/api/bindings/\{provider\}	List bindings for provider
POST	/api/bindings	Create binding
PUT	/api/bindings/\{id\}	Update binding
DELETE	/api/bindings/\{id\}	Delete binding
POST	/api/bindings/resolve	Resolve binding for a message
GET	/api/bindings/stats	Router statistics

Email Services API

Follow-up tracking, snooze, and category APIs live under /api/v1/email.

Method	Endpoint	Description
POST	/api/v1/email/followups/mark	Mark follow-up
GET	/api/v1/email/followups/pending	List pending follow-ups
POST	/api/v1/email/followups/\{id\}/resolve	Resolve follow-up
POST	/api/v1/email/followups/check-replies	Check for replies
POST	/api/v1/email/followups/auto-detect	Auto-detect follow-ups
GET	/api/v1/email/\{id\}/snooze-suggestions	Snooze suggestions
POST	/api/v1/email/\{id\}/snooze	Apply snooze
DELETE	/api/v1/email/\{id\}/snooze	Cancel snooze
GET	/api/v1/email/snoozed	List snoozed emails
GET	/api/v1/email/categories	List categories
POST	/api/v1/email/categories/learn	Category feedback

Gmail Operations API

Gmail labels, threads, drafts, and message operations.

Method	Endpoint	Description
POST	/api/v1/gmail/labels	Create label
GET	/api/v1/gmail/labels	List labels
PATCH	/api/v1/gmail/labels/\{id\}	Update label
DELETE	/api/v1/gmail/labels/\{id\}	Delete label
POST	/api/v1/gmail/messages/\{id\}/labels	Modify message labels
POST	/api/v1/gmail/messages/\{id\}/read	Mark read/unread
POST	/api/v1/gmail/messages/\{id\}/star	Star/unstar
POST	/api/v1/gmail/messages/\{id\}/archive	Archive message
POST	/api/v1/gmail/messages/\{id\}/trash	Trash/untrash
POST	/api/v1/gmail/filters	Create filter
GET	/api/v1/gmail/filters	List filters
DELETE	/api/v1/gmail/filters/\{id\}	Delete filter
GET	/api/v1/gmail/threads	List threads
GET	/api/v1/gmail/threads/\{id\}	Get thread
POST	/api/v1/gmail/threads/\{id\}/archive	Archive thread
POST	/api/v1/gmail/threads/\{id\}/trash	Trash thread
POST	/api/v1/gmail/threads/\{id\}/labels	Modify thread labels
POST	/api/v1/gmail/drafts	Create draft
GET	/api/v1/gmail/drafts	List drafts
GET	/api/v1/gmail/drafts/\{id\}	Get draft
PUT	/api/v1/gmail/drafts/\{id\}	Update draft
DELETE	/api/v1/gmail/drafts/\{id\}	Delete draft
POST	/api/v1/gmail/drafts/\{id\}/send	Send draft
GET	/api/v1/gmail/messages/\{id\}/attachments/\{attachment_id\}	Get attachment

Notes:

• Gmail operations require a connected Gmail account via /api/v1/email/gmail/oauth/*.
• Pass user_id in query params for GET requests and in the JSON body for write requests.

Outlook/M365 API

Method	Endpoint	Description
GET	/api/v1/outlook/oauth/url	OAuth authorization URL
POST	/api/v1/outlook/oauth/callback	OAuth callback
GET	/api/v1/outlook/folders	List mail folders
GET	/api/v1/outlook/messages	List messages
GET	/api/v1/outlook/messages/\{id\}	Get message details
GET	/api/v1/outlook/conversations/\{id\}	Get conversation
POST	/api/v1/outlook/send	Send message
POST	/api/v1/outlook/reply	Reply to message
GET	/api/v1/outlook/search	Search messages
POST	/api/v1/outlook/messages/\{id\}/read	Mark read/unread
POST	/api/v1/outlook/messages/\{id\}/move	Move message
DELETE	/api/v1/outlook/messages/\{id\}	Delete message

Accounting (QuickBooks) API

Accounting endpoints live under /api/accounting for QuickBooks Online integration and financial dashboards.

Method	Endpoint	Description
GET	/api/accounting/status	Connection status + dashboard data
GET	/api/accounting/connect	Start OAuth connection
GET	/api/accounting/callback	OAuth callback
POST	/api/accounting/disconnect	Disconnect QuickBooks
GET	/api/accounting/customers	List customers
GET	/api/accounting/transactions	List transactions
POST	/api/accounting/report	Generate report

Payroll (Gusto) API

Payroll endpoints live under /api/accounting/gusto for Gusto OAuth and payroll data sync.

Method	Endpoint	Description
GET	/api/accounting/gusto/status	Connection status
GET	/api/accounting/gusto/connect	Start OAuth connection
GET	/api/accounting/gusto/callback	OAuth callback
POST	/api/accounting/gusto/disconnect	Disconnect Gusto
GET	/api/accounting/gusto/employees	List employees
GET	/api/accounting/gusto/payrolls	List payroll runs
GET	/api/accounting/gusto/payrolls/\{payroll_id\}	Payroll run details
POST	/api/accounting/gusto/payrolls/\{payroll_id\}/journal-entry	Generate journal entry

Banking (Plaid) API

Banking endpoints live under /api/accounting/plaid for Plaid Link integration and bank account syncing.

Method	Endpoint	Description
POST	/api/accounting/plaid/link/token	Create Link token
POST	/api/accounting/plaid/exchange	Exchange public token
GET	/api/accounting/plaid/status	Connection status
POST	/api/accounting/plaid/disconnect	Disconnect Plaid
GET	/api/accounting/plaid/accounts	List bank accounts
GET	/api/accounting/plaid/transactions	List transactions
POST	/api/accounting/plaid/transactions/sync	Sync transactions
GET	/api/accounting/plaid/balance	Get account balances
GET	/api/accounting/plaid/institutions/search	Search institutions

Xero Accounting API

Xero endpoints live under /api/accounting/xero for Xero OAuth and accounting data sync.

Method	Endpoint	Description
GET	/api/accounting/xero/status	Connection status
GET	/api/accounting/xero/connect	Start OAuth connection
GET	/api/accounting/xero/callback	OAuth callback
POST	/api/accounting/xero/disconnect	Disconnect Xero
GET	/api/accounting/xero/contacts	List contacts
GET	/api/accounting/xero/contacts/\{contact_id\}	Contact details
POST	/api/accounting/xero/contacts	Create contact
GET	/api/accounting/xero/invoices	List invoices
GET	/api/accounting/xero/invoices/\{invoice_id\}	Invoice details
POST	/api/accounting/xero/invoices	Create invoice
GET	/api/accounting/xero/accounts	List chart of accounts
GET	/api/accounting/xero/bank-transactions	List bank transactions
POST	/api/accounting/xero/manual-journals	Create manual journal
GET	/api/accounting/xero/payments	List payments

Cost Visibility API

Cost endpoints live under /api/costs. See docs/COST_VISIBILITY.md for dashboard context.

Method	Endpoint	Description
GET	/api/costs	Cost dashboard summary
GET	/api/costs/breakdown	Cost breakdown
GET	/api/costs/timeline	Usage timeline
GET	/api/costs/alerts	Budget alerts
POST	/api/costs/budget	Set budget limits
POST	/api/costs/alerts/\{alert_id\}/dismiss	Dismiss alert

Chat Knowledge Bridge API

Endpoints for chat + knowledge integration:

Method	Endpoint	Description
POST	/api/v1/chat/knowledge/search	Search knowledge from chat
POST	/api/v1/chat/knowledge/inject	Inject knowledge into conversation
POST	/api/v1/chat/knowledge/store	Store chat as knowledge
GET	/api/v1/chat/knowledge/channel/\{id\}/summary	Channel knowledge summary

Knowledge Mound Governance API

Governance endpoints under /api/v1/knowledge/mound/governance.

Method	Endpoint	Description
POST	/api/v1/knowledge/mound/governance/roles	Create a role
POST	/api/v1/knowledge/mound/governance/roles/assign	Assign role to user
POST	/api/v1/knowledge/mound/governance/roles/revoke	Revoke role from user
GET	/api/v1/knowledge/mound/governance/permissions/\{user_id\}	Get user permissions
POST	/api/v1/knowledge/mound/governance/permissions/check	Check permissions
GET	/api/v1/knowledge/mound/governance/audit	Query audit trail
GET	/api/v1/knowledge/mound/governance/audit/user/\{user_id\}	User activity audit
GET	/api/v1/knowledge/mound/governance/stats	Governance stats

Knowledge Mound Maintenance API

Maintenance endpoints under /api/v1/knowledge/mound.

Method	Endpoint	Description
GET	/api/v1/knowledge/mound/dedup/clusters	Find duplicate clusters
GET	/api/v1/knowledge/mound/dedup/report	Generate dedup report
POST	/api/v1/knowledge/mound/dedup/merge	Merge a duplicate cluster
POST	/api/v1/knowledge/mound/dedup/auto-merge	Auto-merge exact duplicates
POST	/api/v1/knowledge/mound/federation/regions	Register federated region (admin)
GET	/api/v1/knowledge/mound/federation/regions	List federated regions
DELETE	/api/v1/knowledge/mound/federation/regions/:id	Unregister federated region
POST	/api/v1/knowledge/mound/federation/sync/push	Push sync to region
POST	/api/v1/knowledge/mound/federation/sync/pull	Pull sync from region
POST	/api/v1/knowledge/mound/federation/sync/all	Sync with all regions
GET	/api/v1/knowledge/mound/federation/status	Federation health status

Threat Intelligence API

Threat intelligence endpoints live under /api/v1/threat.

Method	Endpoint	Description
POST	/api/v1/threat/url	Scan a URL
POST	/api/v1/threat/urls	Batch scan URLs
GET	/api/v1/threat/ip/\{ip_address\}	IP reputation
POST	/api/v1/threat/ips	Batch IP reputation
GET	/api/v1/threat/hash/\{hash_value\}	File hash lookup
POST	/api/v1/threat/hashes	Batch hash lookup
POST	/api/v1/threat/email	Scan email content
GET	/api/v1/threat/status	Service status

Backups API

Disaster recovery and backup management endpoints under /api/v1/backups.

Method	Endpoint	Description
POST	/api/v1/backups	Create a new backup
GET	/api/v1/backups	List all backups
GET	/api/v1/backups/\{backup_id\}	Get backup details
DELETE	/api/v1/backups/\{backup_id\}	Delete a backup
POST	/api/v1/backups/\{backup_id\}/restore	Restore from backup
POST	/api/v1/backups/\{backup_id\}/verify	Verify backup integrity
GET	/api/v1/backups/status	Get backup service status
POST	/api/v1/backups/schedule	Configure backup schedule
GET	/api/v1/backups/schedule	Get backup schedule

Expenses API

Receipt management and expense categorization under /api/v1/expenses.

Method	Endpoint	Description
POST	/api/v1/expenses/receipts	Upload a receipt
GET	/api/v1/expenses/receipts	List receipts
GET	/api/v1/expenses/receipts/\{receipt_id\}	Get receipt details
POST	/api/v1/expenses/receipts/\{receipt_id\}/categorize	Categorize receipt
POST	/api/v1/expenses/receipts/\{receipt_id\}/approve	Approve expense
POST	/api/v1/expenses/sync	Sync with accounting system
GET	/api/v1/expenses/categories	List expense categories
GET	/api/v1/expenses/summary	Get expense summary

RLM (Recursive Language Models) API

Context compression and programmatic context management under /api/v1/rlm.

Method	Endpoint	Description
POST	/api/v1/rlm/compress	Compress context
POST	/api/v1/rlm/decompress	Decompress context
POST	/api/v1/rlm/query	Query compressed context
GET	/api/v1/rlm/sessions	List active sessions
GET	/api/v1/rlm/sessions/\{session_id\}	Get session details
DELETE	/api/v1/rlm/sessions/\{session_id\}	End session
POST	/api/v1/rlm/stream	Stream compressed output

Unified Inbox API

Multi-provider email management under /api/v1/unified-inbox.

Method	Endpoint	Description
GET	/api/v1/unified-inbox	Get unified inbox messages
GET	/api/v1/unified-inbox/accounts	List connected accounts
POST	/api/v1/unified-inbox/accounts	Connect email account
DELETE	/api/v1/unified-inbox/accounts/\{account_id\}	Disconnect account
GET	/api/v1/unified-inbox/messages/\{message_id\}	Get message details
POST	/api/v1/unified-inbox/messages/\{message_id\}/archive	Archive message
POST	/api/v1/unified-inbox/messages/\{message_id\}/reply	Reply to message
POST	/api/v1/unified-inbox/sync	Sync all accounts
GET	/api/v1/unified-inbox/stats	Get inbox statistics

Feedback API

User feedback and NPS collection under /api/v1/feedback.

Method	Endpoint	Description
POST	/api/v1/feedback/nps	Submit NPS feedback
POST	/api/v1/feedback/general	Submit general feedback
GET	/api/v1/feedback/nps/summary	Get NPS summary (admin)
GET	/api/v1/feedback/prompts	Get feedback prompts
GET	/api/v1/feedback/history	Get feedback history

New Endpoints (2026-02-03)

Endpoint	Description	Status
POST /api/testfixer/analyze	Analyze a failing test	NEW
POST /api/testfixer/propose	Propose a fix via multi-agent debate	NEW
POST /api/testfixer/apply	Apply a proposed fix	NEW
POST /api/testfixer/run	Run the full fix loop	NEW

New Endpoints (2026-01-27)

Endpoint	Description	Status
POST /api/debates/:id/intervention/pause	Pause a debate	NEW
POST /api/debates/:id/intervention/resume	Resume a debate	NEW
POST /api/debates/:id/intervention/inject	Inject argument or follow‑up	NEW
POST /api/debates/:id/intervention/weights	Adjust agent weight	NEW
POST /api/debates/:id/intervention/threshold	Adjust consensus threshold	NEW
GET /api/debates/:id/intervention/state	Get intervention state	NEW
GET /api/debates/:id/intervention/log	Get intervention log	NEW
GET /api/v1/knowledge/mound/dedup/clusters	Find duplicate clusters	NEW
GET /api/v1/knowledge/mound/dedup/report	Generate dedup report	NEW
POST /api/v1/knowledge/mound/dedup/merge	Merge a duplicate cluster	NEW
POST /api/v1/knowledge/mound/dedup/auto-merge	Auto-merge exact duplicates	NEW
POST /api/v1/knowledge/mound/federation/regions	Register federated region	NEW
POST /api/v1/knowledge/mound/federation/sync/push	Push to region	NEW
POST /api/v1/knowledge/mound/federation/sync/pull	Pull from region	NEW
POST /api/v1/knowledge/mound/federation/sync/all	Sync all regions	NEW
GET /api/v1/knowledge/mound/federation/status	Federation status	NEW

New Endpoints (2026-01-22)

Endpoint	Description	Status
POST /api/control-plane/deliberations	Run or queue a vetted decisionmaking session	NEW
GET /api/control-plane/deliberations/:id	Get vetted decisionmaking result	NEW
GET /api/control-plane/deliberations/:id/status	Get vetted decisionmaking status	NEW
POST /api/v1/decisions	Create a decision request	NEW
GET /api/v1/decisions/:id	Get decision result	NEW
GET /api/v1/decisions/:id/status	Get decision status	NEW
POST /api/v1/github/pr/review	Trigger a PR review	NEW
GET /api/v1/github/pr/:pr_number	Get PR details	NEW
GET /api/v1/github/pr/review/:review_id	Get review status	NEW
GET /api/v1/github/pr/:pr_number/reviews	List PR reviews	NEW
POST /api/v1/github/pr/:pr_number/review	Submit review to GitHub	NEW
POST /api/v1/github/audit/issues	Create GitHub issues from findings	NEW
POST /api/v1/chat/knowledge/search	Knowledge search from chat	NEW
POST /api/v1/audit/sessions	Create audit session	NEW
POST /api/v1/inbox/shared	Create shared inbox	NEW
GET /api/v1/inbox/shared	List shared inboxes	NEW
POST /api/v1/inbox/routing/rules	Create routing rule	NEW
GET /api/costs	Cost dashboard summary	NEW
GET /api/accounting/status	Accounting dashboard status	NEW
POST /api/codebase/quick-scan	One-click security scan	NEW

New Endpoints (2026-01-09)

Endpoint	Description	Status
POST /api/debates/graph	Graph-structured debates with branching	NEW
GET /api/debates/graph/:id	Get graph debate by ID	NEW
GET /api/debates/graph/:id/branches	Get branches for graph debate	NEW
GET /api/debates/graph/:id/nodes	Get nodes for graph debate	NEW
POST /api/debates/matrix	Parallel scenario debates	NEW
GET /api/debates/matrix/:id	Get matrix debate by ID	NEW
GET /api/debates/matrix/:id/scenarios	Get scenario results	NEW
GET /api/debates/matrix/:id/conclusions	Get matrix conclusions	NEW
GET /api/breakpoints/pending	List pending human-in-the-loop breakpoints	NEW
GET /api/breakpoints/:id/status	Get breakpoint status	NEW
POST /api/breakpoints/:id/resolve	Resolve breakpoint with human input	NEW
GET /api/introspection/all	Get all agent introspection data	NEW
GET /api/introspection/leaderboard	Agents ranked by reputation	NEW
GET /api/introspection/agents	List available agents	NEW
GET /api/introspection/agents/:name	Get agent introspection	NEW
GET /api/gallery	List public debates	NEW
GET /api/gallery/:id	Get public debate details	NEW
GET /api/gallery/:id/embed	Get embeddable debate summary	NEW
GET /api/billing/plans	List subscription plans	NEW
GET /api/billing/usage	Get current usage	NEW
GET /api/billing/subscription	Get subscription status	NEW
POST /api/billing/checkout	Create Stripe checkout	NEW
POST /api/billing/portal	Create billing portal session	NEW
POST /api/billing/cancel	Cancel subscription	NEW
POST /api/billing/resume	Resume subscription	NEW
POST /api/webhooks/stripe	Handle Stripe webhooks	NEW
GET /api/v1/memory/analytics	Get comprehensive memory tier analytics	NEW
GET /api/v1/memory/analytics/tier/:tier	Get stats for specific tier	NEW
POST /api/v1/memory/analytics/snapshot	Take manual analytics snapshot	NEW
GET /api/evolution/ab-tests	List all A/B tests	NEW
GET /api/evolution/ab-tests/:id	Get specific A/B test	NEW
GET /api/evolution/ab-tests/:agent/active	Get active test for agent	NEW
POST /api/evolution/ab-tests	Create new A/B test	NEW
POST /api/evolution/ab-tests/:id/record	Record debate result	NEW
POST /api/evolution/ab-tests/:id/conclude	Conclude test	NEW
DELETE /api/evolution/ab-tests/:id	Cancel test	NEW

New Endpoints (2026-01-20)

Endpoint	Description	Status
GET /api/v1/debates/:id/explanation	Get full decision explanation	NEW
GET /api/v1/debates/:id/evidence	Get evidence chain	NEW
GET /api/v1/debates/:id/votes/pivots	Get vote pivot analysis	NEW
GET /api/v1/debates/:id/counterfactuals	Get counterfactual scenarios	NEW
GET /api/v1/debates/:id/summary	Get human-readable summary	NEW
GET /api/workflow/templates	List workflow templates	NEW
GET /api/workflow/templates/:id	Get template details	NEW
GET /api/workflow/templates/:id/package	Get full template package	NEW
POST /api/workflow/templates/:id/run	Execute workflow template	NEW
GET /api/workflow/categories	List template categories	NEW
GET /api/workflow/patterns	List workflow patterns	NEW
POST /api/workflow/patterns/:id/instantiate	Create template from pattern	NEW
GET /api/gauntlet/receipts	List gauntlet receipts	NEW
GET /api/gauntlet/receipts/:id	Get receipt details	NEW
GET /api/gauntlet/receipts/:id/verify	Verify receipt integrity	NEW
GET /api/gauntlet/receipts/:id/export	Export receipt (JSON/HTML/MD/SARIF)	NEW
POST /api/gauntlet/receipts/:id/share	Create shareable receipt link	NEW
GET /api/gauntlet/receipts/stats	Get receipt statistics	NEW
POST /api/v1/explainability/batch	Create batch explanation job	NEW
GET /api/v1/explainability/batch/:id/status	Get batch job status	NEW
GET /api/v1/explainability/batch/:id/results	Get batch job results	NEW
POST /api/v1/explainability/compare	Compare explanations across debates	NEW
GET /api/marketplace/templates	Browse template marketplace	NEW
GET /api/marketplace/templates/:id	Get marketplace template details	NEW
POST /api/marketplace/templates	Publish template to marketplace	NEW
POST /api/marketplace/templates/:id/rate	Rate a template	NEW
POST /api/marketplace/templates/:id/review	Review a template	NEW
POST /api/marketplace/templates/:id/import	Import template to workspace	NEW
GET /api/marketplace/featured	Get featured templates	NEW
GET /api/marketplace/trending	Get trending templates	NEW
GET /api/marketplace/categories	Get marketplace categories	NEW

New Endpoints (2026-01-18)

Endpoint	Description	Status
GET /api/auth/oauth/google	Start Google OAuth flow	NEW
GET /api/auth/oauth/google/callback	Google OAuth callback	NEW
GET /api/auth/oauth/github	Start GitHub OAuth flow	NEW
GET /api/auth/oauth/github/callback	GitHub OAuth callback	NEW
POST /api/auth/oauth/link	Link OAuth account to user	NEW
DELETE /api/auth/oauth/unlink	Unlink OAuth account	NEW
GET /api/auth/oauth/providers	List available OAuth providers	NEW
GET /api/user/oauth-providers	Get user's linked providers	NEW
GET /api/workflows	List workflows	NEW
POST /api/workflows	Create workflow	NEW
GET /api/workflows/:id	Get workflow details	NEW
PUT /api/workflows/:id	Update workflow	NEW
DELETE /api/workflows/:id	Delete workflow	NEW
POST /api/workflows/:id/execute	Execute workflow	NEW
GET /api/workflows/:id/versions	Get workflow versions	NEW
GET /api/workflow-templates	List workflow templates	NEW
GET /api/workflow-templates/:id	Get workflow template	NEW
GET /api/workflow-executions	List workflow executions	NEW
GET /api/workflow-executions/:id	Get execution status	NEW
DELETE /api/workflow-executions/:id	Cancel execution	NEW
GET /api/workflow-approvals	List pending approvals	NEW
POST /api/workflow-approvals/:id	Submit approval decision	NEW
POST /api/retention/policies	Create retention policy	NEW
POST /api/retention/policies/:id/execute	Execute retention policy	NEW
GET /api/retention/expiring	Get expiring items	NEW
POST /api/classify	Classify content sensitivity	NEW
GET /api/classify/policy/:level	Get classification policy	NEW
GET /api/audit/entries	Query audit entries	NEW
GET /api/audit/report	Generate compliance report	NEW
GET /api/audit/verify	Verify audit log integrity	NEW

Recently Connected Endpoints

The following endpoints were identified as unused but are now connected:

Endpoint	Component	Status
GET /api/debates	DebateListPanel	✅ Connected
GET /api/agent/\{agent\}/profile	AgentProfileWrapper	✅ Connected
GET /api/agent/compare	AgentComparePanel	✅ Connected
GET /api/agent/\{agent\}/head-to-head/\{opponent\}	AgentProfileWrapper	✅ Connected
GET /api/agent/\{agent\}/network	AgentProfileWrapper	✅ Connected (includes rivals/allies)
GET /api/history/debates	HistoryPanel	✅ Connected (local API fallback)
GET /api/history/summary	HistoryPanel	✅ Connected (local API fallback)
GET /api/history/cycles	HistoryPanel	✅ Connected (local API fallback)
GET /api/history/events	HistoryPanel	✅ Connected (local API fallback)
GET /api/pulse/trending	TrendingTopicsPanel	✅ Connected
GET /api/analytics/disagreements	AnalyticsPanel	✅ Connected

Remaining High-Value Endpoints (Ready to Wire)

Endpoint	Feature	Priority
GET /api/agent/\{agent\}/history	Agent debate history	MEDIUM
GET /api/agent/\{agent\}/rivals	Direct rivals endpoint	LOW
GET /api/agent/\{agent\}/allies	Direct allies endpoint	LOW
GET /api/debates/\{id\}	Individual debate detail view	LOW

Server Configuration

The unified server exposes HTTP on port 8080 and WebSocket on port 8765 by default.

aragora serve --api-port 8080 --ws-port 8765
# Or directly:
python -m aragora.server.unified_server --api-port 8080 --ws-port 8765

Authentication

API requests may include an Authorization header with a bearer token:

Authorization: Bearer <token>

Rate limiting: 60 requests per minute per token (sliding window).

---

OAuth Integration

OAuth providers allow users to authenticate with third-party accounts.

GET /api/auth/oauth/google

Start Google OAuth flow. Redirects user to Google consent screen.

Parameters:

• redirect_url (string, optional): URL to redirect after successful auth (must be in allowlist)

Response: 302 Redirect to Google OAuth consent screen

GET /api/auth/oauth/google/callback

Handle Google OAuth callback after user consent.

Parameters:

• code (string, required): Authorization code from Google
• state (string, required): State parameter for CSRF protection
• error (string, optional): Error code if user denied consent

Response: 302 Redirect to success URL with auth token

GET /api/auth/oauth/github

Start GitHub OAuth flow. Redirects user to GitHub consent screen.

Parameters:

• redirect_url (string, optional): URL to redirect after successful auth

Response: 302 Redirect to GitHub OAuth consent screen

GET /api/auth/oauth/github/callback

Handle GitHub OAuth callback after user consent.

Parameters:

• code (string, required): Authorization code from GitHub
• state (string, required): State parameter for CSRF protection

Response: 302 Redirect to success URL with auth token

POST /api/auth/oauth/link

Link an OAuth provider account to the current authenticated user.

Request Body:

{
  "provider": "google",
  "code": "authorization_code_from_provider"
}

Response:

{
  "success": true,
  "message": "Account linked successfully"
}

DELETE /api/auth/oauth/unlink

Remove OAuth provider link from the current user's account.

Parameters:

• provider (string, required): OAuth provider to unlink (google or github)

Response:

{
  "success": true,
  "message": "Account unlinked successfully"
}

GET /api/auth/oauth/providers

Get list of configured OAuth providers available for authentication.

Response:

{
  "providers": [
    {"id": "google", "name": "Google", "enabled": true},
    {"id": "github", "name": "GitHub", "enabled": true}
  ]
}

GET /api/user/oauth-providers

Get list of OAuth providers linked to the current user's account.

Response:

{
  "linked_providers": [
    {"provider": "google", "email": "user@gmail.com", "linked_at": "2026-01-15T12:00:00Z"}
  ]
}

---

Workflow Management

Workflows define multi-step automated debate and audit processes.

GET /api/workflows

List workflows with optional filtering.

Parameters:

• category (string): Filter by category (debate, analysis, integration, custom)
• tags (string): Filter by tags (comma-separated)
• search (string): Search in workflow name and description
• limit (int, default=50, max=200): Maximum workflows to return
• offset (int, default=0): Pagination offset

Response:

{
  "workflows": [
    {
      "id": "wf-123",
      "name": "Security Audit Workflow",
      "description": "Automated security analysis pipeline",
      "category": "analysis",
      "tags": ["security", "audit"],
      "status": "active",
      "version": 3,
      "created_at": "2026-01-10T10:00:00Z"
    }
  ],
  "total": 25
}

POST /api/workflows

Create a new workflow definition.

Request Body:

{
  "name": "Code Review Pipeline",
  "description": "Automated code review with multi-agent debate",
  "category": "analysis",
  "tags": ["code", "review"],
  "steps": [
    {
      "id": "step-1",
      "type": "debate",
      "config": {
        "topic": "{{input.code_diff}}",
        "agents": ["claude", "gpt-4", "deepseek-r1"],
        "rounds": 3
      }
    },
    {
      "id": "step-2",
      "type": "approval",
      "config": {
        "required_approvers": 1,
        "timeout_hours": 24
      }
    }
  ],
  "transitions": [
    {"from": "step-1", "to": "step-2", "condition": "consensus_reached"}
  ]
}

GET /api/workflows/:workflow_id

Get detailed workflow definition by ID.

Response:

{
  "id": "wf-123",
  "name": "Security Audit Workflow",
  "description": "...",
  "steps": [...],
  "transitions": [...],
  "input_schema": {...},
  "output_schema": {...},
  "version": 3,
  "created_at": "2026-01-10T10:00:00Z",
  "updated_at": "2026-01-15T14:30:00Z"
}

PUT /api/workflows/:workflow_id

Update an existing workflow. Creates a new version.

DELETE /api/workflows/:workflow_id

Delete a workflow definition.

POST /api/workflows/:workflow_id/execute

Start execution of a workflow.

Request Body:

{
  "inputs": {
    "code_diff": "...",
    "repository": "my-repo"
  },
  "async": true
}

Response (async=true):

{
  "execution_id": "exec-456",
  "status": "pending",
  "workflow_id": "wf-123"
}

GET /api/workflows/:workflow_id/versions

Get version history of a workflow.

Parameters:

• limit (int, default=20): Maximum versions to return

GET /api/workflow-templates

List workflow templates for quick start.

Parameters:

• category (string): Filter templates by category

Response:

{
  "templates": [
    {
      "id": "tmpl-security-audit",
      "name": "Security Audit",
      "description": "Comprehensive security analysis",
      "category": "security",
      "complexity": "medium"
    }
  ]
}

GET /api/workflow-templates/:template_id

Get a specific workflow template for use as starting point.

GET /api/workflow-executions

List workflow executions for runtime dashboard.

Parameters:

• workflow_id (string): Filter by workflow ID
• status (string): Filter by status (pending, running, completed, failed, cancelled)
• limit (int, default=50): Maximum executions to return

Response:

{
  "executions": [
    {
      "id": "exec-456",
      "workflow_id": "wf-123",
      "workflow_name": "Security Audit",
      "status": "running",
      "current_step": "step-2",
      "progress": 0.5,
      "started_at": "2026-01-18T10:00:00Z"
    }
  ]
}

GET /api/workflow-executions/:execution_id

Get detailed status of a workflow execution.

Response:

{
  "id": "exec-456",
  "workflow_id": "wf-123",
  "status": "running",
  "current_step": "step-2",
  "step_results": {
    "step-1": {"status": "completed", "output": {...}}
  },
  "context": {...},
  "started_at": "2026-01-18T10:00:00Z"
}

DELETE /api/workflow-executions/:execution_id

Cancel a running workflow execution.

GET /api/workflow-approvals

List workflow steps awaiting human approval.

Response:

{
  "approvals": [
    {
      "id": "approval-789",
      "execution_id": "exec-456",
      "workflow_name": "Security Audit",
      "step_name": "Final Review",
      "context": {...},
      "requested_at": "2026-01-18T11:00:00Z"
    }
  ]
}

POST /api/workflow-approvals/:approval_id

Submit approval decision for a workflow step.

Request Body:

{
  "decision": "approve",
  "comment": "Looks good, proceeding with deployment"
}

Decisions: approve, reject

---

HTTP API Endpoints

Health & Status

GET /api/health

Health check endpoint.

Response:

{
  "status": "healthy",
  "timestamp": "2026-01-04T12:00:00Z"
}

GET /api/nomic/state

Get current nomic loop state.

Response:

{
  "phase": "debate",
  "stage": "executing",
  "cycle": 5,
  "total_tasks": 9,
  "completed_tasks": 3
}

GET /api/nomic/log

Get recent nomic loop log lines.

Parameters:

• lines (int, default=100, max=1000): Number of log lines to return

---

Debates

GET /api/debates

List recent debates.

Parameters:

• limit (int, default=20, max=100): Maximum debates to return

Response:

{
  "debates": [
    {
      "id": "debate-123",
      "topic": "Rate limiter implementation",
      "consensus_reached": true,
      "confidence": 0.85,
      "timestamp": "2026-01-04T12:00:00Z"
    }
  ]
}

GET /api/debates/:slug

Get a specific debate by slug/ID.

POST /api/debate

Start an ad-hoc debate. Rate limited.

Request Body:

{
  "question": "Should we use token bucket or sliding window for rate limiting?",
  "agents": "anthropic-api,openai-api",
  "rounds": 3,
  "consensus": "majority"
}

Parameters:

• question (string, required): Topic/question to debate
• agents (string, default="grok,anthropic-api,openai-api,deepseek-r1"): Comma-separated agent list (max 10)
• rounds (int, default=3): Number of debate rounds
• consensus (string, default="majority"): Consensus method

Available Agent Types:

API (direct):

Type	Default Model	Notes
anthropic-api	claude-opus-4-5-20251101	Anthropic API, streaming
openai-api	gpt-5.2	OpenAI API, streaming
gemini	gemini-3-pro-preview	Google API, streaming
grok	grok-4-latest	xAI API, streaming
mistral-api	mistral-large-2512	Mistral API
codestral	codestral-latest	Mistral code model
ollama	llama3.2	Local Ollama
lm-studio	local-model	Local LM Studio
kimi	moonshot-v1-8k	Moonshot API

OpenRouter:

Type	Default Model	Notes
openrouter	deepseek/deepseek-chat-v3-0324	Model via model parameter
deepseek	deepseek/deepseek-reasoner	DeepSeek R1 (reasoning)
deepseek-r1	deepseek/deepseek-r1	DeepSeek reasoning
llama	meta-llama/llama-3.3-70b-instruct	Llama 3.3 70B
mistral	mistralai/mistral-large-2411	Mistral Large
qwen	qwen/qwen3-max	Qwen3 Max
qwen-max	qwen/qwen3-max	Qwen3 Max
yi	01-ai/yi-large	Yi Large

CLI:

Type	Default Model	Notes
claude	claude-opus-4-5-20251101	Claude CLI
codex	gpt-5.2-codex	Codex CLI
openai	gpt-5.2	OpenAI CLI
gemini-cli	gemini-3-pro-preview	Gemini CLI
grok-cli	grok-4-latest	Grok CLI
qwen-cli	qwen3-coder	Qwen CLI
deepseek-cli	deepseek-v3	DeepSeek CLI
kilocode	gemini-explorer	Codebase explorer

Response:

{
  "debate_id": "debate-abc123",
  "status": "started",
  "message": "Debate started with 2 agents"
}

POST /api/v1/debates/batch

Submit a batch of debates for processing.

Request Body:

{
  "items": [
    {
      "question": "Evaluate rate limiting options",
      "agents": "anthropic-api,openai-api",
      "rounds": 3,
      "consensus": "majority",
      "priority": 5
    }
  ],
  "webhook_url": "https://example.com/webhooks/batch",
  "max_parallel": 5
}

GET /api/v1/debates/batch/{batch_id}/status

Get batch status and item results.

GET /api/v1/debates/batch

List batch requests.

Parameters:

• limit (int, default=50, max=100): Maximum batches to return
• status (string, optional): Filter by status (pending, processing, completed, failed)

GET /api/debates/:id/export/:format

Export a debate in various formats.

Path Parameters:

• id (string, required): Debate slug/ID
• format (string, required): Export format - json, csv, dot, or html

Query Parameters:

• table (string, optional): For CSV format only - summary (default), messages, critiques, votes, or verifications

Response (JSON format):

{
  "artifact_id": "abc123",
  "debate_id": "debate-slug",
  "task": "Rate limiter implementation",
  "consensus_proof": {
    "reached": true,
    "confidence": 0.85,
    "vote_breakdown": {"anthropic-api": true, "gemini": true}
  },
  "agents": ["anthropic-api", "gemini"],
  "rounds": 3,
  "content_hash": "sha256:abcd1234"
}

Response (CSV format): Returns text/csv with debate data table Response (DOT format): Returns text/vnd.graphviz for visualization with GraphViz Response (HTML format): Returns self-contained HTML viewer with interactive graph

POST /api/debates/graph

Run a graph-structured debate with automatic branching.

Request Body:

{
  "task": "Design a distributed caching system",
  "agents": ["anthropic-api", "openai-api"],
  "max_rounds": 5,
  "branch_policy": {
    "min_disagreement": 0.7,
    "max_branches": 3,
    "auto_merge": true,
    "merge_strategy": "synthesis"
  }
}

Response:

{
  "debate_id": "uuid",
  "task": "Design a distributed caching system",
  "graph": {
    "nodes": [...],
    "edges": [...],
    "root_id": "node-123"
  },
  "branches": [...],
  "merge_results": [...],
  "node_count": 15,
  "branch_count": 2
}

Branch Policy Options:

• min_disagreement (float): Threshold to trigger branching (default: 0.7)
• max_branches (int): Maximum concurrent branches (default: 3)
• auto_merge (bool): Automatically merge converging branches (default: true)
• merge_strategy (string): "best_path", "synthesis", "vote", "weighted", "preserve_all"

GET /api/debates/graph/:id

Get a graph debate by ID.

GET /api/debates/graph/:id/branches

Get all branches for a graph debate.

GET /api/debates/graph/:id/nodes

Get all nodes in a graph debate.

POST /api/debates/matrix

Run parallel scenario debates with comparative analysis.

Request Body:

{
  "task": "Design a rate limiter",
  "agents": ["anthropic-api", "openai-api"],
  "scenarios": [
    {
      "name": "High throughput",
      "parameters": {"rps": 10000},
      "constraints": ["Must handle burst traffic"]
    },
    {
      "name": "Low latency",
      "parameters": {"latency_ms": 10},
      "constraints": ["P99 under 10ms"]
    },
    {
      "name": "Baseline",
      "is_baseline": true
    }
  ],
  "max_rounds": 3
}

Response:

{
  "matrix_id": "uuid",
  "task": "Design a rate limiter",
  "scenario_count": 3,
  "results": [
    {
      "scenario_name": "High throughput",
      "parameters": {"rps": 10000},
      "winner": "anthropic-api",
      "final_answer": "...",
      "confidence": 0.85,
      "consensus_reached": true
    }
  ],
  "universal_conclusions": ["All scenarios reached consensus"],
  "conditional_conclusions": [
    {
      "condition": "When High throughput",
      "conclusion": "Use token bucket algorithm",
      "confidence": 0.85
    }
  ],
  "comparison_matrix": {
    "scenarios": ["High throughput", "Low latency", "Baseline"],
    "consensus_rate": 1.0,
    "avg_confidence": 0.82,
    "avg_rounds": 2.5
  }
}

GET /api/debates/matrix/:id

Get matrix debate results by ID.

GET /api/debates/matrix/:id/scenarios

Get all scenario results for a matrix debate.

GET /api/debates/matrix/:id/conclusions

Get conclusions (universal and conditional) for a matrix debate.

Debate Interventions

Mid‑debate controls for pausing, resuming, injecting user input, and adjusting consensus parameters.

POST /api/debates/{debate_id}/intervention/pause

Pause an active debate.

POST /api/debates/{debate_id}/intervention/resume

Resume a paused debate.

POST /api/debates/{debate_id}/intervention/inject

Inject an argument or follow‑up into the next round.

Request Body:

{
  "content": "What about rate limits for burst traffic?",
  "type": "follow_up",
  "source": "user"
}

POST /api/debates/{debate_id}/intervention/weights

Adjust an agent’s influence weight (0.0–2.0).

Request Body:

{
  "agent": "anthropic-api",
  "weight": 1.5
}

POST /api/debates/{debate_id}/intervention/threshold

Adjust consensus threshold (0.5–1.0).

Request Body:

{
  "threshold": 0.85
}

GET /api/debates/{debate_id}/intervention/state

Get the current intervention state (pause, weights, threshold, pending injections).

GET /api/debates/{debate_id}/intervention/log

Get recent intervention log entries.

---

History (Supabase)

GET /api/history/cycles

Get cycle history.

Parameters:

• loop_id (string, optional): Filter by loop ID
• limit (int, default=50, max=200): Maximum cycles to return

GET /api/history/events

Get event history.

Parameters:

• loop_id (string, optional): Filter by loop ID
• limit (int, default=100, max=500): Maximum events to return

GET /api/history/debates

Get debate history.

Parameters:

• loop_id (string, optional): Filter by loop ID
• limit (int, default=50, max=200): Maximum debates to return

GET /api/history/summary

Get summary statistics.

Parameters:

• loop_id (string, optional): Filter by loop ID

---

Leaderboard & ELO

GET /api/leaderboard

Get agent rankings by ELO.

Parameters:

• limit (int, default=20, max=50): Maximum agents to return
• domain (string, optional): Filter by domain

Response:

{
  "rankings": [
    {
      "agent": "anthropic-api",
      "elo": 1523,
      "wins": 45,
      "losses": 12,
      "domain": "general"
    }
  ]
}

GET /api/matches/recent

Get recent ELO matches.

Parameters:

• limit (int, default=10, max=50): Maximum matches to return
• loop_id (string, optional): Filter by loop ID

GET /api/agent/:name/history

Get an agent's match history.

Parameters:

• limit (int, default=30, max=100): Maximum matches to return

---

Insights

GET /api/insights/recent

Get recent debate insights.

Parameters:

• limit (int, default=20, max=100): Maximum insights to return

Response:

{
  "insights": [
    {
      "type": "pattern",
      "content": "Agents prefer incremental implementations",
      "confidence": 0.78,
      "source_debate": "debate-123"
    }
  ]
}

---

Flip Detection

Position reversal detection API for tracking agent consistency.

GET /api/flips/recent

Get recent position flips across all agents.

Parameters:

• limit (int, default=20, max=100): Maximum flips to return

Response:

{
  "flips": [
    {
      "id": "flip-abc123",
      "agent_name": "gemini",
      "original_claim": "X is optimal",
      "new_claim": "Y is optimal",
      "flip_type": "contradiction",
      "similarity_score": 0.82,
      "detected_at": "2026-01-04T12:00:00Z"
    }
  ],
  "count": 15
}

GET /api/flips/summary

Get aggregate flip statistics.

Response:

{
  "total_flips": 42,
  "by_type": {
    "contradiction": 10,
    "refinement": 25,
    "retraction": 5,
    "qualification": 2
  },
  "by_agent": {
    "gemini": 15,
    "anthropic-api": 12,
    "codex": 10,
    "grok": 5
  },
  "recent_24h": 8
}

GET /api/agent/:name/consistency

Get consistency score for an agent.

Response:

{
  "agent_name": "anthropic-api",
  "total_positions": 150,
  "total_flips": 12,
  "consistency_score": 0.92,
  "contradictions": 2,
  "refinements": 8,
  "retractions": 1,
  "qualifications": 1
}

GET /api/agent/:name/flips

Get flips for a specific agent.

Parameters:

• limit (int, default=20, max=100): Maximum flips to return

---

Consensus Memory

Historical consensus data and similarity search.

GET /api/consensus/similar

Find debates similar to a topic.

Parameters:

• topic (string, required): Topic to search for
• limit (int, default=5, max=20): Maximum results to return

Response:

{
  "query": "rate limiting",
  "results": [
    {
      "topic": "Rate limiter design",
      "conclusion": "Token bucket preferred",
      "strength": "strong",
      "confidence": 0.85,
      "similarity": 0.92,
      "agents": ["anthropic-api", "gemini"],
      "dissent_count": 1,
      "timestamp": "2026-01-03T10:00:00Z"
    }
  ],
  "count": 3
}

GET /api/consensus/settled

Get high-confidence settled topics.

Parameters:

• min_confidence (float, default=0.8): Minimum confidence threshold
• limit (int, default=20, max=100): Maximum topics to return

Response:

{
  "min_confidence": 0.8,
  "topics": [
    {
      "topic": "Consensus algorithm choice",
      "conclusion": "Weighted voting preferred",
      "confidence": 0.95,
      "strength": "strong",
      "timestamp": "2026-01-04T08:00:00Z"
    }
  ],
  "count": 15
}

GET /api/consensus/stats

Get consensus memory statistics.

Response:

{
  "total_consensus": 150,
  "total_dissents": 42,
  "by_strength": {
    "strong": 80,
    "moderate": 50,
    "weak": 20
  },
  "by_domain": {
    "general": 100,
    "architecture": 30,
    "security": 20
  },
  "avg_confidence": 0.78
}

GET /api/consensus/dissents

Get dissenting views relevant to a topic.

Parameters:

• topic (string, required): Topic to search for (max 500 chars)
• domain (string, optional): Filter by domain

Response:

{
  "topic": "rate limiting",
  "domain": null,
  "similar_debates": [
    {
      "topic": "Rate limiter design",
      "conclusion": "Token bucket preferred",
      "confidence": 0.85
    }
  ],
  "dissents_by_type": {
    "alternative": 3,
    "concern": 2,
    "objection": 1
  },
  "unacknowledged_dissents": 2
}

GET /api/consensus/domain/:domain

Get consensus history for a domain.

Parameters:

• limit (int, default=50, max=200): Maximum records to return

---

Pulse (Trending Topics)

Real-time trending topic ingestion for dynamic debate generation.

GET /api/pulse/trending

Get trending topics from social media platforms.

Parameters:

• limit (int, default=10, max=50): Maximum topics to return per platform

Response:

{
  "topics": [
    {
      "topic": "AI regulation debate",
      "platform": "twitter",
      "volume": 15000,
      "category": "tech"
    }
  ],
  "count": 5
}

POST /api/pulse/debate-topic

Start a debate on a trending topic.

Request Body:

{
  "topic": "Should AI be regulated?",
  "rounds": 3,
  "consensus_threshold": 0.7
}

Response:

{
  "debate_id": "pulse-1234567890-abc123",
  "topic": "Should AI be regulated?",
  "status": "started"
}

Pulse Scheduler

Automated debate creation from trending topics.

GET /api/pulse/scheduler/status

Get current scheduler state and metrics.

Response:

{
  "state": "running",
  "run_id": "run-1234567890-abc123",
  "config": {
    "poll_interval_seconds": 300,
    "max_debates_per_hour": 6,
    "min_volume_threshold": 100,
    "min_controversy_score": 0.3,
    "dedup_window_hours": 24
  },
  "metrics": {
    "polls_completed": 15,
    "topics_evaluated": 120,
    "debates_created": 8,
    "duplicates_skipped": 5,
    "uptime_seconds": 3600
  }
}

POST /api/pulse/scheduler/start

Start the pulse debate scheduler.

Response:

{
  "status": "started",
  "run_id": "run-1234567890-abc123"
}

POST /api/pulse/scheduler/stop

Stop the pulse debate scheduler.

Response:

{
  "status": "stopped"
}

POST /api/pulse/scheduler/pause

Pause the scheduler (maintains state but stops polling).

Response:

{
  "status": "paused"
}

POST /api/pulse/scheduler/resume

Resume a paused scheduler.

Response:

{
  "status": "running"
}

PATCH /api/pulse/scheduler/config

Update scheduler configuration.

Request Body:

{
  "max_debates_per_hour": 10,
  "min_controversy_score": 0.4
}

Response:

{
  "status": "updated",
  "config": { ... }
}

GET /api/pulse/scheduler/history

Get history of scheduled debates.

Parameters:

• limit (int, default=50): Maximum records to return
• offset (int, default=0): Pagination offset
• platform (string, optional): Filter by platform
• category (string, optional): Filter by category

Response:

{
  "history": [
    {
      "id": "sched-123",
      "topic_text": "AI ethics debate",
      "platform": "hackernews",
      "category": "tech",
      "debate_id": "pulse-456",
      "created_at": 1704067200,
      "consensus_reached": true,
      "confidence": 0.85
    }
  ],
  "total": 100
}

---

Slack Integration

Slack bot integration for debate notifications and commands.

GET /api/integrations/slack/status

Get Slack integration status.

Response:

{
  "enabled": true,
  "signing_secret_configured": true,
  "bot_token_configured": false,
  "webhook_configured": true
}

POST /api/integrations/slack/commands

Handle Slack slash commands (called by Slack).

Supported Commands:

• /aragora help - Show available commands
• /aragora status - Get system status
• /aragora debate "topic" - Start a debate
• /aragora agents - List top agents by ELO

POST /api/integrations/slack/interactive

Handle Slack interactive components (buttons, menus).

POST /api/integrations/slack/events

Handle Slack Events API callbacks.

---

Agent Profile (Combined)

GET /api/agent/:name/profile

Get a combined profile with ELO, persona, consistency, and calibration data.

Response:

{
  "agent": "anthropic-api",
  "ranking": {
    "rating": 1523,
    "recent_matches": 10
  },
  "persona": {
    "type": "analytical",
    "primary_stance": "pragmatic",
    "specializations": ["architecture", "security"],
    "debate_count": 45
  },
  "consistency": {
    "score": 0.92,
    "recent_flips": 2
  },
  "calibration": {
    "brier_score": 0.15,
    "prediction_count": 30
  }
}

---

Agent Relationship Network

Analyze agent relationships, alliances, and rivalries.

GET /api/agent/:name/network

Get complete influence/relationship network for an agent.

Response:

{
  "agent": "anthropic-api",
  "influences": [["gemini", 0.75], ["openai", 0.62]],
  "influenced_by": [["codex", 0.58]],
  "rivals": [["grok", 0.81]],
  "allies": [["gemini", 0.72]]
}

GET /api/agent/:name/rivals

Get top rivals for an agent.

Parameters:

• limit (int, default=5, max=20): Maximum rivals to return

Response:

{
  "agent": "anthropic-api",
  "rivals": [["grok", 0.81], ["openai", 0.65]],
  "count": 2
}

GET /api/agent/:name/allies

Get top allies for an agent.

Parameters:

• limit (int, default=5, max=20): Maximum allies to return

Response:

{
  "agent": "anthropic-api",
  "allies": [["gemini", 0.72], ["codex", 0.55]],
  "count": 2
}

---

Critique Patterns

Retrieve high-impact critique patterns for learning.

GET /api/critiques/patterns

Get critique patterns ranked by success rate.

Parameters:

• limit (int, default=10, max=50): Maximum patterns to return
• min_success (float, default=0.5): Minimum success rate threshold

Response:

{
  "patterns": [
    {
      "issue_type": "security",
      "pattern": "Consider input validation",
      "success_rate": 0.85,
      "usage_count": 12
    }
  ],
  "count": 5,
  "stats": {
    "total_critiques": 150,
    "total_patterns": 42
  }
}

---

Replays

GET /api/replays

List available debate replays.

Response:

{
  "replays": [
    {
      "id": "nomic-cycle-1",
      "name": "Nomic Cycle 1",
      "event_count": 245,
      "created_at": "2026-01-03T10:00:00Z"
    }
  ]
}

GET /api/replays/:id

Get a specific replay by ID.

---

Broadcast (Podcast Generation)

Generate audio podcasts from debate traces.

POST /api/debates/:id/broadcast

Generate an MP3 podcast from a debate.

Rate limited. Requires the broadcast module (pip install aragora[broadcast]).

Response:

{
  "success": true,
  "debate_id": "rate-limiter-2026-01-01",
  "audio_path": "/tmp/aragora_debate_rate-limiter.mp3",
  "format": "mp3"
}

Error Response (503):

{
  "error": "Broadcast module not available"
}

---

Documents

GET /api/documents

List uploaded documents.

POST /api/documents/upload

Upload a document for processing.

Headers:

• Content-Type: multipart/form-data
• X-Filename: document.pdf

Supported formats: PDF, Markdown, Python, JavaScript, TypeScript, Jupyter notebooks

Max size: 10MB

GET /api/documents/formats

Get supported document formats and metadata.

Response:

{
  "formats": {
    ".pdf": {"name": "PDF", "mime": "application/pdf"},
    ".md": {"name": "Markdown", "mime": "text/markdown"},
    ".py": {"name": "Python", "mime": "text/x-python"},
    ".js": {"name": "JavaScript", "mime": "text/javascript"},
    ".ts": {"name": "TypeScript", "mime": "text/typescript"},
    ".ipynb": {"name": "Jupyter Notebook", "mime": "application/x-ipynb+json"}
  }
}

---

Debate Analytics

Real-time debate analytics and pattern detection.

GET /api/analytics/disagreements

Get debates with significant disagreements or failed consensus.

Parameters:

• limit (int, default=20, max=100): Maximum records to return

Response:

{
  "disagreements": [
    {
      "debate_id": "debate-123",
      "topic": "Rate limiter design",
      "agents": ["anthropic-api", "gemini"],
      "dissent_count": 2,
      "consensus_reached": false,
      "confidence": 0.45,
      "timestamp": "2026-01-04T12:00:00Z"
    }
  ],
  "count": 5
}

GET /api/analytics/role-rotation

Get agent role assignments across debates.

Parameters:

• limit (int, default=50, max=200): Maximum rotations to return

Response:

{
  "rotations": [
    {
      "debate_id": "debate-123",
      "agent": "anthropic-api",
      "role": "proposer",
      "timestamp": "2026-01-04T12:00:00Z"
    }
  ],
  "summary": {
    "anthropic-api": {"proposer": 10, "critic": 8, "judge": 5},
    "gemini": {"proposer": 8, "critic": 12, "judge": 3}
  },
  "count": 50
}

GET /api/analytics/early-stops

Get debates that terminated before completing all planned rounds.

Parameters:

• limit (int, default=20, max=100): Maximum records to return

Response:

{
  "early_stops": [
    {
      "debate_id": "debate-123",
      "topic": "Consensus algorithm",
      "rounds_completed": 2,
      "rounds_planned": 5,
      "reason": "early_consensus",
      "consensus_early": true,
      "timestamp": "2026-01-04T12:00:00Z"
    }
  ],
  "count": 3
}

---

Learning Evolution

GET /api/learning/evolution

Get learning evolution patterns.

---

Modes

List available debate and operational modes.

GET /api/modes

Get all available modes.

Response:

{
  "modes": [
    {
      "name": "architect",
      "description": "High-level design and planning mode",
      "category": "operational",
      "tool_groups": ["read", "browser", "mcp"]
    },
    {
      "name": "redteam",
      "description": "Adversarial red-teaming for security analysis",
      "category": "debate"
    }
  ],
  "count": 8
}

---

Agent Position Tracking

Track agent positions and consistency via truth-grounding system.

GET /api/agent/:name/positions

Get position history for an agent.

Parameters:

• limit (int, default=50, max=200): Maximum positions to return

Response:

{
  "agent": "anthropic-api",
  "total_positions": 45,
  "avg_confidence": 0.82,
  "reversal_count": 3,
  "consistency_score": 0.93,
  "positions_by_debate": {
    "rate-limiter-2026-01-01": 5,
    "consensus-algo-2026-01-02": 3
  }
}

---

System Statistics

System-wide metrics and health monitoring.

GET /api/ranking/stats

Get ELO ranking system statistics.

Response:

{
  "total_agents": 8,
  "total_matches": 245,
  "avg_rating": 1523,
  "rating_spread": 312,
  "domains": ["general", "architecture", "security"],
  "active_last_24h": 5
}

GET /api/v1/memory/tier-stats

Get memory tier statistics from continuum memory system.

Response:

{
  "tiers": {
    "fast": {"count": 50, "avg_importance": 0.85},
    "slow": {"count": 200, "avg_importance": 0.65},
    "glacial": {"count": 1000, "avg_importance": 0.35}
  },
  "total_entries": 1250,
  "last_consolidation": "2026-01-04T12:00:00Z"
}

GET /api/v1/memory/analytics

Get comprehensive memory tier analytics.

Parameters:

• days (int, default=30, min=1, max=365): Number of days to analyze

Response:

{
  "total_memories": 1000,
  "tier_stats": {
    "fast": {"count": 100, "avg_importance": 0.8},
    "medium": {"count": 300, "avg_importance": 0.6},
    "slow": {"count": 400, "avg_importance": 0.4},
    "glacial": {"count": 200, "avg_importance": 0.2}
  },
  "learning_velocity": 0.75,
  "promotion_effectiveness": 0.82
}

GET /api/v1/memory/analytics/tier/:tier

Get stats for a specific memory tier.

Path Parameters:

• tier (string, required): Tier name - fast, medium, slow, or glacial

Parameters:

• days (int, default=30, min=1, max=365): Number of days to analyze

Response:

{
  "tier": "fast",
  "count": 100,
  "avg_importance": 0.8,
  "hit_rate": 0.95,
  "promotion_rate": 0.3
}

POST /api/v1/memory/analytics/snapshot

Take a manual analytics snapshot for all memory tiers.

Response:

{
  "status": "success",
  "message": "Snapshot recorded for all tiers"
}

---

GET /api/critiques/patterns

Get successful critique patterns for learning.

Parameters:

• limit (int, default=10, max=50): Maximum patterns to return
• min_success (float, default=0.5): Minimum success rate filter

Response:

{
  "patterns": [
    {
      "issue_type": "edge_case",
      "pattern": "Consider boundary conditions for...",
      "success_rate": 0.85,
      "usage_count": 23
    }
  ],
  "count": 5,
  "stats": {"total_patterns": 150, "avg_success_rate": 0.72}
}

GET /api/critiques/archive

Get archive statistics for resolved patterns.

Response:

{
  "archived": 42,
  "by_type": {
    "security": 15,
    "performance": 12,
    "edge_case": 15
  }
}

---

Agent Reputation

Track agent reliability and voting weights.

GET /api/reputation/all

Get all agent reputations ranked by score.

Response:

{
  "reputations": [
    {
      "agent": "anthropic-api",
      "score": 0.85,
      "vote_weight": 1.35,
      "proposal_acceptance_rate": 0.78,
      "critique_value": 0.82,
      "debates_participated": 45
    }
  ],
  "count": 4
}

GET /api/agent/:name/reputation

Get reputation for a specific agent.

Response:

{
  "agent": "anthropic-api",
  "score": 0.85,
  "vote_weight": 1.35,
  "proposal_acceptance_rate": 0.78,
  "critique_value": 0.82,
  "debates_participated": 45
}

---

Agent Comparison

Compare agents head-to-head.

GET /api/agent/compare

Get head-to-head comparison between two agents.

Parameters:

• agent_a (string, required): First agent name
• agent_b (string, required): Second agent name

Response:

{
  "agent_a": "anthropic-api",
  "agent_b": "gemini",
  "matches": 15,
  "agent_a_wins": 9,
  "agent_b_wins": 6,
  "win_rate_a": 0.6,
  "domains": {
    "architecture": {"a_wins": 5, "b_wins": 2},
    "security": {"a_wins": 4, "b_wins": 4}
  }
}

---

Agent Expertise & Grounded Personas

Evidence-based agent identity and expertise tracking.

GET /api/agent/:name/domains

Get agent's best expertise domains by calibration score.

Parameters:

• limit (int, default=5, max=20): Maximum domains to return

Response:

{
  "agent": "anthropic-api",
  "domains": [
    {"domain": "security", "calibration_score": 0.89},
    {"domain": "api_design", "calibration_score": 0.85}
  ],
  "count": 2
}

GET /api/agent/:name/grounded-persona

Get truth-grounded persona synthesized from performance data.

Response:

{
  "agent": "anthropic-api",
  "elo": 1523,
  "domain_elos": {"security": 1580, "architecture": 1490},
  "games_played": 45,
  "win_rate": 0.62,
  "calibration_score": 0.78,
  "position_accuracy": 0.72,
  "positions_taken": 128,
  "reversals": 8
}

GET /api/agent/:name/identity-prompt

Get evidence-grounded identity prompt for agent initialization.

Parameters:

• sections (string, optional): Comma-separated sections to include (performance, calibration, relationships, positions)

Response:

{
  "agent": "anthropic-api",
  "identity_prompt": "## Your Identity: anthropic-api\nYour approach: analytical, thorough...",
  "sections": ["performance", "calibration"]
}

---

Contrarian Views & Risk Warnings

Historical dissenting views and edge case concerns from past debates.

GET /api/consensus/contrarian-views

Get historical contrarian views on a topic.

Parameters:

• topic (string, required): Topic to search for contrarian views
• domain (string, optional): Filter by domain
• limit (int, default=5, max=20): Maximum views to return

Response:

{
  "topic": "rate limiting implementation",
  "domain": null,
  "contrarian_views": [
    {
      "agent": "gemini",
      "position": "Token bucket has edge cases",
      "reasoning": "Under burst traffic conditions...",
      "confidence": 0.75,
      "timestamp": "2026-01-03T10:00:00Z"
    }
  ],
  "count": 1
}

GET /api/consensus/risk-warnings

Get risk warnings and edge case concerns from past debates.

Parameters:

• topic (string, required): Topic to search for risk warnings
• domain (string, optional): Filter by domain
• limit (int, default=5, max=20): Maximum warnings to return

Response:

{
  "topic": "database migration",
  "domain": "infrastructure",
  "risk_warnings": [
    {
      "agent": "anthropic-api",
      "warning": "Consider rollback strategy for schema changes",
      "severity": "high",
      "timestamp": "2026-01-02T15:00:00Z"
    }
  ],
  "count": 1
}

---

Head-to-Head & Opponent Analysis

Detailed comparison and strategic briefings between agents.

GET /api/agent/:agent/head-to-head/:opponent

Get detailed head-to-head statistics between two agents.

Response:

{
  "agent": "anthropic-api",
  "opponent": "gemini",
  "matches": 12,
  "agent_wins": 5,
  "opponent_wins": 4,
  "draws": 3,
  "win_rate": 0.42,
  "recent_form": "WDLWW"
}

GET /api/agent/:agent/opponent-briefing/:opponent

Get strategic briefing about an opponent for an agent.

Response:

{
  "agent": "anthropic-api",
  "opponent": "gemini",
  "briefing": {
    "relationship": "rival",
    "strength": 0.7,
    "head_to_head": {"wins": 5, "losses": 4, "draws": 3},
    "opponent_strengths": ["visual reasoning", "synthesis"],
    "opponent_weaknesses": ["consistency", "edge cases"],
    "recommended_strategy": "Focus on logical rigor and consistency"
  }
}

---

Calibration Analysis

Detailed calibration curves and prediction accuracy.

GET /api/agent/:name/calibration-curve

Get calibration curve showing expected vs actual accuracy per confidence bucket.

Parameters:

• buckets (int, default=10, max=20): Number of confidence buckets
• domain (string, optional): Filter by domain

Response:

{
  "agent": "anthropic-api",
  "domain": null,
  "buckets": [
    {
      "range_start": 0.0,
      "range_end": 0.1,
      "total_predictions": 15,
      "correct_predictions": 2,
      "accuracy": 0.13,
      "expected_accuracy": 0.05,
      "brier_score": 0.08
    },
    {
      "range_start": 0.9,
      "range_end": 1.0,
      "total_predictions": 42,
      "correct_predictions": 38,
      "accuracy": 0.90,
      "expected_accuracy": 0.95,
      "brier_score": 0.02
    }
  ],
  "count": 10
}

---

Meta-Critique Analysis

Analyze debate quality and identify process issues.

GET /api/debate/:id/meta-critique

Get meta-level analysis of a debate including repetition, circular arguments, and ignored critiques.

Response:

{
  "debate_id": "debate-123",
  "overall_quality": 0.72,
  "productive_rounds": [1, 2, 4],
  "unproductive_rounds": [3],
  "observations": [
    {
      "type": "repetition",
      "severity": "medium",
      "agent": "gemini",
      "round": 3,
      "description": "Agent repeated similar points from round 1"
    }
  ],
  "recommendations": [
    "Encourage agents to address critiques directly",
    "Consider reducing round count for simple topics"
  ]
}

---

Agent Personas

Agent persona definitions and customizations.

GET /api/personas

Get all agent personas.

Response:

{
  "personas": [
    {
      "agent_name": "anthropic-api",
      "description": "Analytical reasoner focused on logical consistency",
      "traits": ["analytical", "precise", "evidence-focused"],
      "expertise": ["security", "architecture"],
      "created_at": "2026-01-01T00:00:00Z"
    },
    {
      "agent_name": "grok",
      "description": "Creative problem-solver with lateral thinking",
      "traits": ["creative", "unconventional", "synthesis-focused"],
      "expertise": ["innovation", "edge-cases"],
      "created_at": "2026-01-01T00:00:00Z"
    }
  ],
  "count": 2
}

---

Persona Laboratory

Experimental framework for evolving agent personas and detecting emergent traits.

GET /api/laboratory/emergent-traits

Get emergent traits detected from agent performance patterns.

Parameters:

• min_confidence (float, default=0.5): Minimum confidence threshold
• limit (int, default=10, max=50): Maximum traits to return

Response:

{
  "emergent_traits": [
    {
      "agent": "anthropic-api",
      "trait": "adversarial_robustness",
      "domain": "security",
      "confidence": 0.85,
      "evidence": "Consistently identifies edge cases in security debates",
      "detected_at": "2026-01-04T10:00:00Z"
    }
  ],
  "count": 1,
  "min_confidence": 0.5
}

POST /api/laboratory/cross-pollinations/suggest

Suggest beneficial trait transfers for a target agent.

Request Body:

{
  "target_agent": "gemini"
}

Response:

{
  "target_agent": "gemini",
  "suggestions": [
    {
      "source_agent": "anthropic-api",
      "trait_or_domain": "logical_rigor",
      "reason": "Target agent underperforms in formal reasoning domains"
    }
  ],
  "count": 1
}

---

Belief Network Analysis

Bayesian belief network for probabilistic debate reasoning.

GET /api/belief-network/:debate_id/cruxes

Identify key claims that would most impact the debate outcome.

Parameters:

• top_k (int, default=3, max=10): Number of cruxes to return

Response:

{
  "debate_id": "debate-123",
  "cruxes": [
    {
      "claim_id": "claim-456",
      "statement": "The proposed architecture scales linearly",
      "score": 0.87,
      "centrality": 0.9,
      "uncertainty": 0.75
    }
  ],
  "count": 1
}

---

Provenance & Evidence Chain

Verify evidence provenance and claim support.

GET /api/provenance/:debate_id/claims/:claim_id/support

Get verification status of all evidence supporting a claim.

Response:

{
  "debate_id": "debate-123",
  "claim_id": "claim-456",
  "support": {
    "verified": true,
    "evidence_count": 3,
    "supporting": [
      {
        "evidence_id": "ev-001",
        "type": "citation",
        "integrity_verified": true,
        "relevance": 0.92
      }
    ],
    "contradicting": []
  }
}

---

Agent Routing & Selection

Optimal agent selection for tasks based on ELO, expertise, and team dynamics.

POST /api/routing/recommendations

Get agent recommendations for a task.

Request Body:

{
  "task_id": "design-review",
  "primary_domain": "architecture",
  "secondary_domains": ["security", "performance"],
  "required_traits": ["analytical"],
  "limit": 5
}

Response:

{
  "task_id": "design-review",
  "primary_domain": "architecture",
  "recommendations": [
    {
      "name": "anthropic-api",
      "type": "anthropic-api",
      "match_score": 0.92,
      "domain_expertise": 0.85
    }
  ],
  "count": 1
}

---

Tournament System

Tournament management and standings.

GET /api/tournaments/:tournament_id/standings

Get current tournament standings.

Response:

{
  "tournament_id": "round-robin-2026",
  "standings": [
    {
      "agent": "anthropic-api",
      "wins": 8,
      "losses": 2,
      "draws": 1,
      "points": 25,
      "total_score": 142.5,
      "win_rate": 0.73
    }
  ],
  "count": 4
}

---

Team Analytics

Analyze team performance and find optimal combinations.

GET /api/routing/best-teams

Get best-performing team combinations from history.

Parameters:

• min_debates (int, default=3, max=20): Minimum debates for a team to qualify
• limit (int, default=10, max=50): Maximum combinations to return

Response:

{
  "min_debates": 3,
  "combinations": [
    {
      "agents": ["anthropic-api", "gemini"],
      "success_rate": 0.85,
      "total_debates": 12,
      "wins": 10
    }
  ],
  "count": 5
}

---

Prompt Evolution

Track agent prompt evolution and learning.

GET /api/evolution/:agent/history

Get prompt evolution history for an agent.

Parameters:

• limit (int, default=10, max=50): Maximum history entries to return

Response:

{
  "agent": "anthropic-api",
  "history": [
    {
      "from_version": 1,
      "to_version": 2,
      "strategy": "pattern_mining",
      "patterns_applied": ["logical_rigor", "edge_case_handling"],
      "created_at": "2026-01-04T08:00:00Z"
    }
  ],
  "count": 3
}

---

Evolution A/B Testing

Run controlled experiments comparing prompt versions to determine which performs better.

GET /api/evolution/ab-tests

List all A/B tests with optional filters.

Parameters:

• agent (string, optional): Filter by agent name
• status (string, optional): Filter by status (active, concluded, cancelled)
• limit (int, default=50, max=200): Maximum tests to return

Response:

{
  "tests": [
    {
      "id": "test-123",
      "agent": "anthropic-api",
      "baseline_prompt_version": 1,
      "evolved_prompt_version": 2,
      "status": "active",
      "baseline_wins": 5,
      "evolved_wins": 7,
      "started_at": "2026-01-10T12:00:00Z"
    }
  ],
  "count": 1
}

GET /api/evolution/ab-tests/:id

Get a specific A/B test by ID.

Response:

{
  "id": "test-123",
  "agent": "anthropic-api",
  "baseline_prompt_version": 1,
  "evolved_prompt_version": 2,
  "status": "active",
  "baseline_wins": 5,
  "evolved_wins": 7,
  "baseline_debates": 10,
  "evolved_debates": 10,
  "evolved_win_rate": 0.58,
  "is_significant": false,
  "started_at": "2026-01-10T12:00:00Z",
  "concluded_at": null
}

GET /api/evolution/ab-tests/:agent/active

Get the active A/B test for a specific agent.

Response:

{
  "agent": "anthropic-api",
  "has_active_test": true,
  "test": {
    "id": "test-123",
    "baseline_prompt_version": 1,
    "evolved_prompt_version": 2,
    "status": "active"
  }
}

POST /api/evolution/ab-tests

Create a new A/B test.

Request Body:

{
  "agent": "anthropic-api",
  "baseline_version": 1,
  "evolved_version": 2,
  "metadata": {"description": "Test new reasoning patterns"}
}

Response (201 Created):

{
  "message": "A/B test created",
  "test": {
    "id": "test-456",
    "agent": "anthropic-api",
    "status": "active"
  }
}

Error (409 Conflict): Agent already has an active test.

POST /api/evolution/ab-tests/:id/record

Record a debate result for an A/B test.

Request Body:

{
  "debate_id": "debate-789",
  "variant": "evolved",
  "won": true
}

Response:

{
  "message": "Result recorded",
  "test": {
    "id": "test-123",
    "baseline_wins": 5,
    "evolved_wins": 8
  }
}

POST /api/evolution/ab-tests/:id/conclude

Conclude an A/B test and determine the winner.

Request Body:

{
  "force": false
}

Response:

{
  "message": "A/B test concluded",
  "result": {
    "test_id": "test-123",
    "winner": "evolved",
    "confidence": 0.85,
    "recommendation": "Adopt evolved prompt",
    "stats": {
      "evolved_win_rate": 0.65,
      "baseline_win_rate": 0.35,
      "total_debates": 20
    }
  }
}

DELETE /api/evolution/ab-tests/:id

Cancel an active A/B test.

Response:

{
  "message": "A/B test cancelled",
  "test_id": "test-123"
}

---

Load-Bearing Claims

Identify claims with highest structural importance in debates.

GET /api/belief-network/:debate_id/load-bearing-claims

Get claims with highest centrality (most load-bearing).

Parameters:

• limit (int, default=5, max=20): Maximum claims to return

Response:

{
  "debate_id": "debate-123",
  "load_bearing_claims": [
    {
      "claim_id": "claim-456",
      "statement": "The architecture must support horizontal scaling",
      "author": "anthropic-api",
      "centrality": 0.92
    }
  ],
  "count": 3
}

---

Calibration Summary

Comprehensive agent calibration analysis.

GET /api/agent/:name/calibration-summary

Get comprehensive calibration summary for an agent.

Parameters:

• domain (string, optional): Filter by domain

Response:

{
  "agent": "anthropic-api",
  "domain": null,
  "total_predictions": 250,
  "total_correct": 215,
  "accuracy": 0.86,
  "brier_score": 0.12,
  "ece": 0.05,
  "is_overconfident": false,
  "is_underconfident": true
}

GET /api/calibration/leaderboard

Get agents ranked by calibration quality (accuracy vs confidence).

Parameters:

• limit (int, default=10, max=50): Maximum agents to return
• domain (string, optional): Filter by domain

Response:

{
  "agents": [
    {
      "name": "anthropic-api",
      "elo": 1542,
      "calibration_score": 0.92,
      "brier_score": 0.08,
      "accuracy": 0.89,
      "games": 45
    },
    {
      "name": "grok",
      "elo": 1518,
      "calibration_score": 0.88,
      "brier_score": 0.11,
      "accuracy": 0.85,
      "games": 38
    }
  ],
  "count": 2
}

---

Continuum Memory

Multi-timescale memory system with surprise-weighted importance scoring.

Note: /api/memory/* legacy routes are supported as aliases for /api/v1/memory/*.

GET /api/v1/memory/continuum/retrieve

Retrieve memories from the continuum memory system.

Parameters:

• query (string, optional): Search query for memory content
• tiers (string, default="fast,medium"): Comma-separated tier names (fast, medium, slow, glacial)
• limit (int, default=10, max=50): Maximum memories to return
• min_importance (float, default=0.0): Minimum importance threshold (0.0-1.0)

Response:

{
  "query": "error patterns",
  "tiers": ["FAST", "MEDIUM"],
  "memories": [
    {
      "id": "mem-123",
      "tier": "FAST",
      "content": "TypeError pattern in agent responses",
      "importance": 0.85,
      "surprise_score": 0.3,
      "consolidation_score": 0.7,
      "success_rate": 0.92,
      "update_count": 15,
      "created_at": "2026-01-03T10:00:00Z",
      "updated_at": "2026-01-04T08:00:00Z"
    }
  ],
  "count": 1
}

POST /api/v1/memory/continuum/consolidate

Run memory consolidation and get tier transition statistics.

Response:

{
  "consolidation": {
    "promoted": 5,
    "demoted": 2,
    "pruned": 1
  },
  "message": "Memory consolidation complete"
}

GET /api/v1/memory/search-index

Progressive disclosure stage 1: compact index results with previews and token estimates.

Parameters:

• q (string, required): Search query
• tier (string, optional): Comma-separated tier filter
• limit (int, default=20, max=100): Maximum results
• min_importance (float, default=0.0): Minimum importance threshold
• use_hybrid (bool, default=false): Use hybrid vector+keyword search
• include_external (bool, default=false): Include external memory sources
• external (string, optional): Comma-separated sources (supermemory, claude-mem)
• project (string, optional): claude-mem project filter

GET /api/v1/memory/search-timeline

Progressive disclosure stage 2: timeline around an anchor entry.

Parameters:

• anchor_id (string, required): Memory ID to anchor timeline
• before (int, default=3, max=50): Entries before anchor
• after (int, default=3, max=50): Entries after anchor
• tier (string, optional): Comma-separated tier filter
• min_importance (float, default=0.0): Minimum importance threshold

GET /api/v1/memory/entries

Progressive disclosure stage 3: full entries by ID.

Parameters:

• ids (string, required): Comma-separated memory IDs

GET /api/v1/memory/viewer

HTML Memory Viewer UI (uses progressive disclosure endpoints).

---

Formal Verification

POST /api/verification/formal-verify

Attempt formal verification of a claim using Z3 SMT solver.

Request Body:

{
  "claim": "If X > Y and Y > Z, then X > Z",
  "claim_type": "logical",
  "context": "Transitivity property of greater-than",
  "timeout": 30
}

Parameters:

• claim (required): The claim to verify
• claim_type (optional): Hint for the type (assertion, logical, arithmetic, constraint)
• context (optional): Additional context for translation
• timeout (optional): Timeout in seconds (default: 30, max: 120)

Response:

{
  "claim": "If X > Y and Y > Z, then X > Z",
  "status": "proof_found",
  "is_verified": true,
  "language": "z3_smt",
  "formal_statement": "(declare-const X Int)...",
  "proof_hash": "a1b2c3d4e5f6",
  "proof_search_time_ms": 15.2,
  "prover_version": "z3-4.12.0"
}

Status Values:

• proof_found: Claim is formally verified
• proof_failed: Counterexample found (claim is false)
• translation_failed: Could not translate to formal language
• timeout: Solver timed out
• not_supported: Claim type not suitable for formal proof
• backend_unavailable: Z3 not installed

---

Insight Extraction

POST /api/insights/extract-detailed

Extract detailed insights from debate content.

Request Body:

{
  "content": "The debate transcript content...",
  "debate_id": "debate-123",
  "extract_claims": true,
  "extract_evidence": true,
  "extract_patterns": true
}

Response:

{
  "debate_id": "debate-123",
  "content_length": 5420,
  "claims": [
    {
      "text": "Therefore, we should adopt this approach",
      "position": 12,
      "type": "argument"
    }
  ],
  "evidence_chains": [
    {
      "text": "According to the 2024 study",
      "type": "citation",
      "source": "the 2024 study"
    }
  ],
  "patterns": [
    {
      "type": "causal_reasoning",
      "strength": "strong",
      "instances": 5
    }
  ]
}

Pattern Types:

• balanced_comparison: Uses "on one hand... on the other hand"
• concession_rebuttal: Uses "while... however"
• enumerated_argument: Uses "first... second..."
• conditional_reasoning: Uses "if... then"
• causal_reasoning: Uses "because" (with instance count)

---

Capability Probing

POST /api/probes/run

Run capability probes against an agent to detect vulnerabilities.

Request Body:

{
  "agent": "anthropic-api",
  "strategies": ["contradiction", "hallucination"],
  "probe_count": 3
}

Response:

{
  "agent": "anthropic-api",
  "probe_count": 3,
  "available_strategies": ["contradiction", "hallucination", "sycophancy", "persistence"],
  "status": "ready"
}

---

Red Team Analysis

POST /api/debates/:id/red-team

Run adversarial analysis on a debate's conclusions.

Request Body:

{
  "attack_types": ["steelman", "strawman"],
  "intensity": 5
}

Response:

{
  "debate_id": "debate-123",
  "task": "Security implementation",
  "consensus_reached": true,
  "intensity": 5,
  "available_attacks": ["steelman", "strawman", "edge_case", "assumption_probe", "counterexample"],
  "status": "ready"
}

---

Usage Examples

Common API operations with curl. Replace localhost:8080 with your server address.

Health Check

# Check server health
curl http://localhost:8080/api/health

Starting a Debate

# Start a new debate
curl -X POST http://localhost:8080/api/debates/start \
  -H "Content-Type: application/json" \
  -d '{
    "question": "Is microservices architecture better than monolith?",
    "agents": ["anthropic-api", "openai-api", "gemini"],
    "rounds": 3
  }'

Listing and Viewing Debates

# List recent debates (default limit: 20)
curl http://localhost:8080/api/debates

# List with custom limit
curl "http://localhost:8080/api/debates?limit=50"

# Get specific debate details
curl http://localhost:8080/api/debates/debate-abc123

Exporting Debates

# Export as JSON
curl http://localhost:8080/api/debates/debate-abc123/export?format=json

# Export messages as CSV
curl http://localhost:8080/api/debates/debate-abc123/export?format=csv&table=messages \
  -o debate-messages.csv

# Export as standalone HTML
curl http://localhost:8080/api/debates/debate-abc123/export?format=html \
  -o debate-report.html

Agent Information

# Get leaderboard
curl http://localhost:8080/api/agent/leaderboard

# Get agent profile
curl http://localhost:8080/api/agent/anthropic-api/profile

# Compare two agents
curl "http://localhost:8080/api/agent/compare?a=anthropic-api&b=openai-api"

# Get head-to-head record
curl http://localhost:8080/api/agent/anthropic-api/head-to-head/openai-api

Nomic Loop Status

# Get current nomic loop state
curl http://localhost:8080/api/nomic/state

# Get nomic loop logs (last 100 lines)
curl http://localhost:8080/api/nomic/log

# Get more log lines
curl "http://localhost:8080/api/nomic/log?lines=500"

Tournaments

# List tournaments
curl http://localhost:8080/api/tournaments

# Get tournament details
curl http://localhost:8080/api/tournaments/tourney-abc123

# Get tournament bracket
curl http://localhost:8080/api/tournaments/tourney-abc123/bracket

Authenticated Requests

# With bearer token
curl http://localhost:8080/api/debates \
  -H "Authorization: Bearer your-token-here"

# Tokens in query parameters are not accepted; use Authorization header instead

WebSocket Connection (wscat)

# Connect to WebSocket for real-time events
wscat -c ws://localhost:8765/ws

# With authentication token
wscat -c ws://localhost:8765/ws -H "Authorization: Bearer your-token-here"

---

WebSocket API

Connect to the WebSocket server for real-time streaming:

const ws = new WebSocket('ws://localhost:8765/ws');

ws.onmessage = (event) => {
  const data = JSON.parse(event.data);
  console.log(data.type, data.data);
};

Event Types

WebSocket events use a shared envelope and are documented in docs/WEBSOCKET_EVENTS.md. Filter by loop_id to scope to a single debate.

Common debate lifecycle events:

• debate_start, round_start, agent_message, critique, vote, consensus, debate_end

Token streaming events:

• token_start, token_delta, token_end

Control messages (on connect / acknowledgements):

• connection_info, loop_list, sync, ack, error, auth_revoked

Event Format

{
  "type": "agent_message",
  "data": { "content": "I disagree...", "role": "critic" },
  "timestamp": 1732735053.123,
  "round": 2,
  "agent": "anthropic-api",
  "loop_id": "loop-abc123",
  "seq": 42
}

---

Formal Verification

Formal claim verification using theorem provers.

POST /api/verification/formal-verify

Attempt formal verification of a logical claim.

Request Body:

{
  "claim": "If A implies B and B implies C, then A implies C",
  "debate_id": "debate-123",
  "context": "Optional context for the claim"
}

Response:

{
  "verified": true,
  "method": "lean",
  "proof_sketch": "...",
  "confidence": 0.95
}

---

Detailed Insights

POST /api/insights/extract-detailed

Extract detailed insights from debate content.

Request Body:

{
  "debate_id": "debate-123",
  "content": "The debate transcript or summary...",
  "focus": "security"
}

Response:

{
  "insights": [
    {
      "type": "pattern",
      "description": "Security vulnerability pattern detected",
      "confidence": 0.85,
      "actionable": true
    }
  ],
  "count": 3
}

---

Formal Verification

Attempt formal verification of claims using SMT solvers.

GET /api/verification/status

Get status of formal verification backends.

Response:

{
  "available": true,
  "backends": [
    {"language": "z3_smt", "available": true},
    {"language": "lean4", "available": false}
  ]
}

POST /api/verification/formal-verify

Attempt formal verification of a claim.

Request Body:

{
  "claim": "If X > Y and Y > Z, then X > Z",
  "claim_type": "logical",
  "context": "Transitivity check",
  "timeout": 30
}

Response:

{
  "status": "proof_found",
  "is_verified": true,
  "language": "z3_smt",
  "formal_statement": "(assert (not (=> (and (> x y) (> y z)) (> x z))))",
  "proof_hash": "a1b2c3d4e5f6",
  "proof_search_time_ms": 15.4
}

---

Analytics & Patterns

Analyze debate patterns and agent behavior.

GET /api/analytics/disagreements

Get analysis of agent disagreement patterns.

Parameters:

• limit (int, default=20, max=100): Maximum entries

Response:

{
  "patterns": [
    {
      "topic": "Error handling strategies",
      "agents": ["anthropic-api", "gemini"],
      "disagreement_rate": 0.73,
      "debates_count": 8
    }
  ]
}

GET /api/analytics/role-rotation

Get role rotation analysis across debates.

Parameters:

• limit (int, default=50, max=200): Maximum entries

GET /api/analytics/early-stops

Get early termination signals and patterns.

Parameters:

• limit (int, default=20, max=100): Maximum entries

---

Agent Network & Relationships

Analyze agent relationships based on debate history.

GET /api/agent/:name/network

Get agent's relationship network with other agents.

Response:

{
  "agent": "anthropic-api",
  "connections": [
    {"agent": "gemini", "relationship": "rival", "strength": 0.8},
    {"agent": "deepseek", "relationship": "ally", "strength": 0.65}
  ]
}

GET /api/agent/:name/rivals

Get agent's top rivals (agents they often disagree with).

Parameters:

• limit (int, default=5, max=20): Maximum rivals

GET /api/agent/:name/allies

Get agent's top allies (agents they often agree with).

Parameters:

• limit (int, default=5, max=20): Maximum allies

GET /api/agent/:name/positions

Get agent's historical positions on topics.

Parameters:

• limit (int, default=50, max=200): Maximum positions

---

Laboratory & Emergent Traits

Track emergent agent behaviors and traits.

GET /api/laboratory/emergent-traits

Get emergent traits discovered across agents.

Parameters:

• min_confidence (float, default=0.5): Minimum confidence threshold
• limit (int, default=10, max=50): Maximum traits

Response:

{
  "traits": [
    {
      "trait": "contrarian_stance",
      "agents": ["grok"],
      "confidence": 0.82,
      "evidence_count": 15
    }
  ]
}

POST /api/laboratory/cross-pollinations/suggest

Suggest trait cross-pollination between agents.

---

Belief Network

Analyze claim dependencies and belief propagation.

GET /api/belief-network/:debate_id/cruxes

Get debate cruxes (key disagreement points).

Parameters:

• top_k (int, default=3, max=10): Number of cruxes to return

Response:

{
  "debate_id": "debate-123",
  "cruxes": [
    {
      "claim": "Performance matters more than readability",
      "centrality": 0.85,
      "agents_for": ["gemini"],
      "agents_against": ["anthropic-api"]
    }
  ]
}

GET /api/belief-network/:debate_id/load-bearing-claims

Get claims that most influence the debate outcome.

Parameters:

• limit (int, default=5, max=20): Maximum claims

---

Tournaments

Agent tournament management.

GET /api/tournaments

List tournaments.

Response:

{
  "tournaments": [
    {
      "id": "t-001",
      "name": "Weekly Championship",
      "status": "in_progress",
      "participants": 8
    }
  ]
}

GET /api/tournaments/:id/standings

Get tournament standings and bracket.

---

Modes

Available stress-test and interaction modes.

GET /api/modes

List available modes.

Response:

{
  "modes": [
    {"name": "standard", "description": "Standard decision stress-test (multi-agent debate engine)"},
    {"name": "adversarial", "description": "Red-team adversarial mode"},
    {"name": "consensus", "description": "Consensus-building mode"}
  ]
}

---

Routing & Team Selection

Agent selection and team composition.

GET /api/routing/best-teams

Get best performing agent team combinations.

Parameters:

• min_debates (int, default=3, max=20): Minimum debates together
• limit (int, default=10, max=50): Maximum teams

Response:

{
  "teams": [
    {
      "agents": ["anthropic-api", "gemini", "deepseek"],
      "win_rate": 0.85,
      "debates": 12,
      "avg_confidence": 0.82
    }
  ]
}

POST /api/routing/recommendations

Get agent routing recommendations for a topic.

Request Body:

{
  "topic": "API security design",
  "required_count": 3
}

---

Error Responses

All endpoints return errors in this format:

{
  "error": "Description of the error"
}

Common HTTP status codes:

• 400 - Bad request (invalid parameters)
• 403 - Forbidden (access denied)
• 404 - Not found
• 500 - Internal server error

---

CORS Policy

The API allows cross-origin requests from:

• http://localhost:3000
• http://localhost:8080
• https://aragora.ai
• https://www.aragora.ai

Other origins are blocked by the browser.

---

Rate Limits

Endpoint	Limit
General API	60 req/min per token
Document upload	10 req/min
WebSocket	Unlimited messages

---

Security Notes

1. Path traversal protection: All file paths are validated to prevent directory traversal attacks
2. Input validation: All integer/float parameters have bounds checking
3. Error sanitization: API keys and tokens are redacted from error messages
4. Origin validation: CORS uses allowlist instead of wildcard
5. SQL injection prevention: LIKE patterns are escaped to prevent wildcard injection
6. Rate limiting: Expensive endpoints (debate creation, uploads) are rate limited
7. Query bounds: Maximum 10 agents per debate, 10 parts per multipart upload
8. Database timeouts: SQLite connections have 30-second timeout to prevent deadlocks
9. Content-Length validation: Headers validated to prevent integer parsing attacks

---

Python Module APIs

The following sections document the Python APIs for extending Aragora.

Plugin System (aragora.plugins)

The plugin system provides a sandboxed execution environment for code analysis, linting, and security scanning extensions.

PluginManifest

Defines plugin metadata and capabilities.

from aragora.plugins.manifest import PluginManifest, PluginCapability, PluginRequirement

manifest = PluginManifest(
    name="my-linter",
    version="1.0.0",
    description="Custom code linter",
    entry_point="my_linter:run",  # module:function format
    capabilities=[PluginCapability.LINT, PluginCapability.CODE_ANALYSIS],
    requirements=[PluginRequirement.READ_FILES],
    timeout_seconds=30.0,
    tags=["python", "lint"],
)

# Validate manifest
valid, errors = manifest.validate()

PluginCapability Enum:

• CODE_ANALYSIS - Analyze code structure
• LINT - Check code style/issues
• SECURITY_SCAN - Security vulnerability detection
• FORMAT - Code formatting
• TEST - Test execution

PluginRequirement Enum:

• READ_FILES - Read file access
• WRITE_FILES - Write file access
• NETWORK - Network access
• SUBPROCESS - Subprocess execution

PluginRunner

Executes plugins in a sandboxed environment with restricted builtins.

from aragora.plugins.runner import PluginRunner, PluginContext

runner = PluginRunner(manifest)

# Create execution context
ctx = PluginContext(
    working_dir="/path/to/project",
    input_data={"files": ["main.py", "utils.py"]},
)
ctx.allowed_operations = {"read_files"}

# Run plugin
result = await runner.run(ctx, timeout_override=60.0)

if result.success:
    print(result.output)
else:
    print(result.errors)

Security Features:

• Restricted builtins (no exec, eval, compile, __import__, open)
• Path traversal prevention (cannot access parent directories)
• Configurable timeouts
• Capability-based permissions

PluginRegistry

Manages plugin discovery and instantiation.

from aragora.plugins.runner import get_registry, run_plugin

# Get global registry
registry = get_registry()

# List all plugins
plugins = registry.list_plugins()

# Get plugins by capability
lint_plugins = registry.list_by_capability(PluginCapability.LINT)

# Run a plugin by name
result = await run_plugin("lint", {"files": ["main.py"]})

---

Genesis System (aragora.genesis)

The Genesis system implements evolutionary agent algorithms with genetic operators and provenance tracking.

AgentGenome

Represents an agent's genetic makeup with traits, expertise, and fitness tracking.

from aragora.genesis import AgentGenome, generate_genome_id

# Create a genome
genome = AgentGenome(
    genome_id=generate_genome_id(
        traits={"analytical": 0.8, "creative": 0.6},
        expertise={"security": 0.9},
        parents=[]
    ),
    name="security-specialist-v1",
    traits={"analytical": 0.8, "creative": 0.6},
    expertise={"security": 0.9, "backend": 0.7},
    model_preference="anthropic-api",
    generation=0,
    fitness_score=0.5,
)

# Create from existing Persona
genome = AgentGenome.from_persona(persona, model="gemini")

# Convert back to Persona for debate use
persona = genome.to_persona()

# Update fitness based on debate outcome
genome.update_fitness(
    consensus_win=True,
    critique_accepted=True,
    prediction_correct=False
)

# Calculate genetic similarity (0-1)
similarity = genome1.similarity_to(genome2)

GenomeStore

SQLite-based persistence for genomes.

from aragora.genesis import GenomeStore

store = GenomeStore(db_path=".nomic/genesis.db")

# Save genome
store.save(genome)

# Retrieve by ID
genome = store.get("abc123def456")

# Get by name (returns latest generation)
genome = store.get_by_name("security-specialist")

# Get top performers
top_10 = store.get_top_by_fitness(n=10)

# Get lineage (ancestors)
lineage = store.get_lineage(genome_id)  # [child, parent, grandparent, ...]

GenomeBreeder

Genetic operators for evolving agent populations.

from aragora.genesis import GenomeBreeder

breeder = GenomeBreeder(
    mutation_rate=0.1,      # Probability of mutating each trait
    crossover_ratio=0.5,    # Blend ratio (0.5 = equal parents)
    elite_ratio=0.2,        # Fraction preserved unchanged
)

# Crossover: blend two parents
child = breeder.crossover(
    parent_a=genome1,
    parent_b=genome2,
    name="hybrid-agent",
    debate_id="debate-123"
)

# Mutation: random modifications
mutated = breeder.mutate(genome, rate=0.2)

# Spawn specialist: create domain-focused agent
specialist = breeder.spawn_specialist(
    domain="security",
    parent_pool=[genome1, genome2, genome3],
    debate_id="debate-123"
)

Population

Collection of genomes with aggregate statistics.

from aragora.genesis import Population

pop = Population(
    population_id="gen-5",
    genomes=[genome1, genome2, genome3],
    generation=5,
)

print(f"Size: {pop.size}")
print(f"Average fitness: {pop.average_fitness:.2f}")
print(f"Best: {pop.best_genome.name}")

# Find by ID
genome = pop.get_by_id("abc123")

GenesisLedger

Immutable event log with cryptographic hashing.

from aragora.genesis import GenesisLedger, GenesisEvent, GenesisEventType

ledger = GenesisLedger(db_path=".nomic/genesis.db")

# Events are automatically hashed and linked
# Common event types:
# - DEBATE_START, DEBATE_END, DEBATE_SPAWN, DEBATE_MERGE
# - CONSENSUS_REACHED, TENSION_DETECTED, TENSION_RESOLVED
# - AGENT_BIRTH, AGENT_DEATH, AGENT_MUTATION, AGENT_CROSSOVER
# - FITNESS_UPDATE, POPULATION_EVOLVED, GENERATION_ADVANCE

FractalTree

Tree structure for nested sub-debates.

from aragora.genesis import FractalTree

tree = FractalTree(root_id="main-debate")
tree.add_node("main-debate", parent_id=None, depth=0, success=True)
tree.add_node("sub-1", parent_id="main-debate", tension="scope", depth=1)
tree.add_node("sub-2", parent_id="main-debate", tension="definition", depth=1)

children = tree.get_children("main-debate")  # ["sub-1", "sub-2"]
nested_dict = tree.to_dict()  # Recursive structure

---

Evolution System (aragora.evolution)

Prompt evolution enables agents to improve their system prompts based on successful debate patterns.

PromptEvolver

Main class for evolving agent prompts.

from aragora.evolution import PromptEvolver, EvolutionStrategy

evolver = PromptEvolver(
    db_path="aragora_evolution.db",
    strategy=EvolutionStrategy.HYBRID,
)

# Extract patterns from successful debates
patterns = evolver.extract_winning_patterns(
    debates=recent_debates,
    min_confidence=0.6,
)

# Store patterns for future use
evolver.store_patterns(patterns)

# Get most effective patterns
top_patterns = evolver.get_top_patterns(
    pattern_type="issue_identification",
    limit=10,
)

# Evolve an agent's prompt
new_prompt = evolver.evolve_prompt(
    agent=my_agent,
    patterns=top_patterns,
    strategy=EvolutionStrategy.APPEND,
)

# Apply evolution and save version
evolver.apply_evolution(agent, patterns)

# Track performance
evolver.update_performance(
    agent_name="anthropic-api",
    version=3,
    debate_result=result,
)

EvolutionStrategy

Available strategies for prompt evolution:

from aragora.evolution import EvolutionStrategy

# APPEND: Add new learnings section to existing prompt
# REPLACE: Replace old learnings section with new patterns
# REFINE: Use LLM to synthesize patterns into coherent prompt
# HYBRID: Append first, refine if prompt exceeds 2000 chars

Prompt Versioning

Track prompt versions and their performance:

# Save a new version
version = evolver.save_prompt_version(
    agent_name="anthropic-api",
    prompt="You are a helpful assistant...",
    metadata={"source": "manual", "note": "Added security focus"},
)

# Get specific version
v1 = evolver.get_prompt_version("anthropic-api", version=1)

# Get latest version
latest = evolver.get_prompt_version("anthropic-api")

# Get evolution history
history = evolver.get_evolution_history("anthropic-api", limit=10)
# Returns: [{"from_version": 2, "to_version": 3, "strategy": "append", ...}, ...]

---

Breakpoints API

Human-in-the-loop breakpoint management for debate supervision.

List Pending Breakpoints

GET /api/breakpoints/pending

Returns all breakpoints awaiting human resolution.

Response:

{
  "breakpoints": [
    {
      "id": "bp_abc123",
      "debate_id": "debate_xyz",
      "type": "consensus_uncertain",
      "created_at": "2026-01-09T10:00:00Z",
      "context": {
        "agents_involved": ["anthropic-api", "openai-api"],
        "disagreement_score": 0.85
      }
    }
  ]
}

Get Breakpoint Status

GET /api/breakpoints/:id/status

Returns the current status of a specific breakpoint.

Resolve Breakpoint

POST /api/breakpoints/:id/resolve

Resolve a pending breakpoint with human guidance.

Request Body:

{
  "resolution": "continue",
  "guidance": "The agents should focus on practical implementation",
  "override_winner": null
}

---

Introspection API

Agent self-awareness and reputation metrics.

Get All Agent Introspection

GET /api/introspection/all

Returns introspection data for all agents.

Get Introspection Leaderboard

GET /api/introspection/leaderboard

Returns agents ranked by reputation score.

Response:

{
  "rankings": [
    {
      "agent": "anthropic-api",
      "reputation_score": 0.92,
      "consistency": 0.88,
      "win_rate": 0.67
    }
  ]
}

List Available Agents

GET /api/introspection/agents

Returns list of agents available for introspection.

Get Agent Availability

GET /api/introspection/agents/availability

Returns credential availability for known agent types.

Get Agent Introspection

GET /api/introspection/agents/:name

Returns detailed introspection for a specific agent.

---

Gallery API

Public debate gallery for sharing and embedding.

List Public Debates

GET /api/gallery

Query Parameters:

Parameter	Type	Default	Description
limit	int	20	Max debates to return
offset	int	0	Pagination offset
agent	string	-	Filter by agent name

Response:

{
  "debates": [
    {
      "id": "gallery_abc123",
      "title": "Should AI systems be open-sourced?",
      "agents": ["anthropic-api", "openai-api"],
      "winner": "anthropic-api",
      "created_at": "2026-01-09T10:00:00Z",
      "view_count": 1234
    }
  ],
  "total": 156,
  "has_more": true
}

Get Public Debate

GET /api/gallery/:id

Returns full debate history for a public debate.

Get Embeddable Summary

GET /api/gallery/:id/embed

Returns an embeddable summary suitable for sharing.

Response:

{
  "embed_html": "<div class='aragora-embed'>...</div>",
  "og_tags": {
    "title": "AI Debate: Open Source AI",
    "description": "Claude vs GPT-4 debate on open-source AI",
    "image": "https://api.aragora.ai/og/gallery_abc123.png"
  }
}

---

Billing API

Subscription and usage management (requires authentication).

List Plans

GET /api/billing/plans

Returns available subscription plans with features and pricing.

Response:

{
  "plans": [
    {
      "id": "free",
      "name": "Free",
      "price": 0,
      "debates_per_month": 10,
      "features": ["basic_agents", "public_gallery"]
    },
    {
      "id": "pro",
      "name": "Pro",
      "price": 29,
      "debates_per_month": 500,
      "features": ["all_agents", "private_debates", "api_access"]
    }
  ]
}

Get Current Usage

GET /api/billing/usage

Returns current usage for the authenticated user.

Response:

{
  "debates_used": 45,
  "debates_limit": 500,
  "api_calls_used": 1234,
  "api_calls_limit": 10000,
  "period_ends": "2026-02-01T00:00:00Z"
}

Get Subscription Status

GET /api/billing/subscription

Returns current subscription details.

Create Checkout Session

POST /api/billing/checkout

Create a Stripe checkout session for plan upgrade.

Request Body:

{
  "plan_id": "pro",
  "success_url": "https://aragora.ai/billing/success",
  "cancel_url": "https://aragora.ai/billing/cancel"
}

Response:

{
  "checkout_url": "https://checkout.stripe.com/..."
}

Create Billing Portal Session

POST /api/billing/portal

Create a Stripe billing portal session for subscription management.

Cancel Subscription

POST /api/billing/cancel

Cancel subscription at end of current billing period.

Resume Subscription

POST /api/billing/resume

Resume a previously canceled subscription.

Stripe Webhook

POST /api/webhooks/stripe

Handle Stripe webhook events (subscription updates, payment events).

---

Usage Metering API

Enterprise usage metering endpoints for tracking consumption and quotas.

Get Usage Summary

GET /api/v1/billing/usage
GET /api/v1/billing/usage/summary
Authorization: Bearer <token>

Query Parameters:

• period: Billing period (hour, day, week, month, quarter, year). Default: month

Returns current usage for the authenticated user's organization.

Response:

{
  "usage": {
    "period_start": "2026-01-01T00:00:00Z",
    "period_end": "2026-01-31T23:59:59Z",
    "period_type": "month",
    "tokens": {
      "input": 500000,
      "output": 250000,
      "total": 750000,
      "cost": "12.50"
    },
    "counts": {
      "debates": 45,
      "api_calls": 1500
    },
    "by_provider": {...},
    "limits": {...},
    "usage_percent": {...}
  }
}

Get Usage Breakdown

GET /api/v1/billing/usage/breakdown
Authorization: Bearer <token>

Query Parameters:

• start: Start date (ISO format)
• end: End date (ISO format)

Returns detailed usage breakdown by model, provider, and day.

Response:

{
  "breakdown": {
    "totals": {
      "cost": "125.50",
      "tokens": 5000000,
      "debates": 150,
      "api_calls": 5000
    },
    "by_model": [...],
    "by_provider": [...],
    "by_day": [...],
    "by_user": [...]
  }
}

Get Limits

GET /api/v1/billing/limits
Authorization: Bearer <token>

Returns current tier limits and utilization percentages.

Response:

{
  "limits": {
    "tier": "enterprise_plus",
    "limits": {
      "tokens": 999999999,
      "debates": 999999,
      "api_calls": 999999
    },
    "used": {
      "tokens": 750000,
      "debates": 45,
      "api_calls": 1500
    },
    "percent": {
      "tokens": 0.075,
      "debates": 0.0045,
      "api_calls": 0.15
    },
    "exceeded": {
      "tokens": false,
      "debates": false,
      "api_calls": false
    }
  }
}

Get Quota Status

GET /api/v1/quotas
Authorization: Bearer <token>

Returns current quota status for all resources using the unified QuotaManager.

Response:

{
  "quotas": {
    "debates": {
      "limit": 100,
      "current": 45,
      "remaining": 55,
      "period": "day",
      "percentage_used": 45.0,
      "is_exceeded": false,
      "is_warning": false,
      "resets_at": "2026-01-24T00:00:00Z"
    },
    "api_requests": {...},
    "tokens": {...},
    "storage_bytes": {...},
    "knowledge_bytes": {...}
  }
}

Export Usage

GET /api/v1/billing/usage/export
Authorization: Bearer <token>

Query Parameters:

• start: Start date (ISO format)
• end: End date (ISO format)
• format: Export format (csv or json). Default: csv

Returns usage data as downloadable CSV or JSON.

---

SME Usage Dashboard API

Comprehensive usage visibility and ROI tracking for SME users.

Get Usage Summary

GET /api/v1/usage/summary
Authorization: Bearer <token>

Query Parameters:

• period: Time period (hour, day, week, month, quarter, year). Default: month
• start: Custom start date (ISO format)
• end: Custom end date (ISO format)

Returns unified usage metrics including debates, costs, tokens, and activity.

Response:

{
  "summary": {
    "period": {
      "type": "month",
      "start": "2025-12-27T00:00:00Z",
      "end": "2026-01-26T23:59:59Z",
      "days": 30
    },
    "debates": {
      "total": 45,
      "completed": 40,
      "consensus_rate": 85.5
    },
    "costs": {
      "total_usd": "12.50",
      "avg_per_debate_usd": "0.28",
      "by_provider": {
        "anthropic": "8.50",
        "openai": "4.00"
      }
    },
    "tokens": {
      "total": 750000,
      "input": 500000,
      "output": 250000
    },
    "activity": {
      "active_days": 15,
      "debates_per_day": 1.5,
      "api_calls": 1500
    }
  }
}

Get Usage Breakdown

GET /api/v1/usage/breakdown
Authorization: Bearer <token>

Query Parameters:

• dimension: Breakdown dimension (agent, model, day, debate). Default: agent
• period: Time period
• start: Custom start date
• end: Custom end date

Returns detailed usage breakdown by the specified dimension.

Response:

{
  "breakdown": {
    "dimension": "agent",
    "period": {
      "start": "2025-12-27T00:00:00Z",
      "end": "2026-01-26T23:59:59Z"
    },
    "items": [
      {
        "name": "claude",
        "cost_usd": "8.50",
        "percentage": 68.0
      },
      {
        "name": "gpt-4",
        "cost_usd": "4.00",
        "percentage": 32.0
      }
    ]
  }
}

Get ROI Metrics

GET /api/v1/usage/roi
Authorization: Bearer <token>

Query Parameters:

• benchmark: Industry benchmark (sme, enterprise, tech_startup, consulting). Default: sme
• period: Time period
• hourly_rate: Override hourly rate assumption (USD)

Returns ROI analysis comparing AI-assisted decisions with traditional methods.

Response:

{
  "roi": {
    "period": {
      "start": "2025-12-27T00:00:00Z",
      "end": "2026-01-26T23:59:59Z"
    },
    "time_savings": {
      "total_hours_saved": 45.0,
      "avg_hours_per_debate": 1.5,
      "value_of_time_saved_usd": "3375.00"
    },
    "cost": {
      "total_ai_cost_usd": "12.50",
      "avg_cost_per_debate_usd": "0.28",
      "traditional_cost_estimate_usd": "3500.00"
    },
    "roi": {
      "percentage": 26900,
      "net_savings_usd": "3362.50",
      "payback_period_debates": 1
    },
    "quality": {
      "consensus_rate": 85.5,
      "avg_confidence": 0.87,
      "decisions_with_high_confidence": 38
    },
    "productivity": {
      "debates_per_active_day": 3.0,
      "time_per_decision_minutes": 5.2,
      "traditional_time_per_decision_hours": 4.0
    },
    "benchmark": {
      "name": "sme",
      "hourly_rate_usd": "75.00",
      "avg_decision_cost_usd": "300.00",
      "avg_hours_per_decision": 4.0
    }
  }
}

Get Budget Status

GET /api/v1/usage/budget-status
Authorization: Bearer <token>

Returns current budget utilization and projections.

Response:

{
  "budget": {
    "monthly": {
      "limit_usd": "100.00",
      "spent_usd": "45.50",
      "remaining_usd": "54.50",
      "percent_used": 45.5,
      "days_remaining": 15,
      "projected_end_spend_usd": "91.00"
    },
    "daily": {
      "limit_usd": "10.00",
      "spent_usd": "1.50"
    },
    "alert_level": null
  }
}

Get Usage Forecast

GET /api/v1/usage/forecast
Authorization: Bearer <token>

Query Parameters:

• benchmark: Industry benchmark for ROI projections

Returns usage forecast based on current patterns.

Response:

{
  "forecast": {
    "projected_monthly_debates": 60,
    "projected_monthly_cost_usd": "16.80",
    "projected_time_savings_hours": 60.0,
    "projected_roi_percentage": 28000,
    "confidence": 0.85,
    "recommendations": [
      "Current usage pattern is efficient",
      "Consider increasing debate usage for higher ROI"
    ]
  }
}

Get Industry Benchmarks

GET /api/v1/usage/benchmarks
Authorization: Bearer <token>

Returns industry benchmark comparison data.

Response:

{
  "benchmarks": {
    "sme": {
      "name": "Small/Medium Enterprise",
      "hourly_rate_usd": "75.00",
      "avg_decision_cost_usd": "300.00",
      "avg_hours_per_decision": 4.0,
      "avg_participants": 3
    },
    "enterprise": {
      "name": "Enterprise",
      "hourly_rate_usd": "150.00",
      "avg_decision_cost_usd": "1200.00",
      "avg_hours_per_decision": 8.0,
      "avg_participants": 5
    },
    "tech_startup": {
      "name": "Tech Startup",
      "hourly_rate_usd": "100.00",
      "avg_decision_cost_usd": "400.00",
      "avg_hours_per_decision": 4.0,
      "avg_participants": 4
    },
    "consulting": {
      "name": "Consulting",
      "hourly_rate_usd": "250.00",
      "avg_decision_cost_usd": "2000.00",
      "avg_hours_per_decision": 8.0,
      "avg_participants": 4
    }
  }
}

Export Usage Data

GET /api/v1/usage/export
Authorization: Bearer <token>

Query Parameters:

• format: Export format (csv, json, pdf). Default: csv
• period: Time period
• start: Custom start date
• end: Custom end date
• include_roi: Include ROI metrics (true/false). Default: false

Returns usage data as downloadable file.

Response (JSON format):

{
  "organization": "My Company",
  "period": {
    "start": "2025-12-27T00:00:00Z",
    "end": "2026-01-26T23:59:59Z"
  },
  "totals": {
    "cost_usd": "12.50",
    "tokens": 750000,
    "api_calls": 1500
  },
  "by_agent": {
    "claude": "8.50",
    "gpt-4": "4.00"
  },
  "by_model": {
    "claude-3-opus": "5.00",
    "claude-3-sonnet": "3.50",
    "gpt-4-turbo": "4.00"
  },
  "roi": {
    "time_saved_hours": 45.0,
    "net_savings_usd": "3362.50"
  }
}

---

Job Queue API

Background job queue management for long-running tasks.

Submit Job

POST /api/queue/jobs
Authorization: Bearer <token>
Content-Type: application/json

{
  "type": "debate_export",
  "payload": {
    "debate_id": "debate-123",
    "format": "json"
  },
  "priority": "normal"
}

Response:

{
  "job_id": "job-abc123",
  "status": "pending",
  "position": 3
}

List Jobs

GET /api/queue/jobs?status=pending&limit=20

Query parameters:

• status: Filter by status (pending, running, completed, failed)
• type: Filter by job type
• limit: Max results (default 50)
• offset: Pagination offset

Get Job Status

GET /api/queue/jobs/\{job_id\}

Response:

{
  "job_id": "job-abc123",
  "type": "debate_export",
  "status": "running",
  "progress": 45,
  "created_at": "2026-01-18T10:00:00Z",
  "started_at": "2026-01-18T10:00:05Z"
}

Retry Failed Job

POST /api/queue/jobs/\{job_id\}/retry
Authorization: Bearer <token>

Cancel Job

DELETE /api/queue/jobs/\{job_id\}
Authorization: Bearer <token>

Queue Statistics

GET /api/queue/stats

Response:

{
  "pending": 12,
  "running": 3,
  "completed_24h": 156,
  "failed_24h": 2,
  "avg_wait_time_ms": 1250,
  "avg_processing_time_ms": 4500
}

Worker Status

GET /api/queue/workers

Response:

{
  "workers": [
    {
      "id": "worker-1",
      "status": "busy",
      "current_job": "job-abc123",
      "jobs_completed": 42
    }
  ]
}

---

Moments API

Track and query significant debate moments (upsets, reversals, breakthroughs).

Get Moments Summary

GET /api/moments/summary

Response:

{
  "total_moments": 156,
  "by_type": {
    "upset_victory": 23,
    "position_reversal": 18,
    "consensus_breakthrough": 45,
    "calibration_vindication": 12,
    "alliance_shift": 8,
    "streak_achievement": 30,
    "domain_mastery": 20
  },
  "recent_count_24h": 12
}

Get Moments Timeline

GET /api/moments/timeline?limit=50&offset=0

Response:

{
  "moments": [
    {
      "id": "moment-123",
      "type": "upset_victory",
      "debate_id": "debate-456",
      "agent": "claude-3",
      "description": "Won against higher-rated opponent",
      "significance": 0.85,
      "timestamp": "2026-01-18T10:30:00Z"
    }
  ],
  "total": 156
}

Get Trending Moments

GET /api/moments/trending?limit=10

Returns most significant recent moments.

Get Moments by Type

GET /api/moments/by-type/\{type\}

Valid types: upset_victory, position_reversal, calibration_vindication, alliance_shift, consensus_breakthrough, streak_achievement, domain_mastery

---

Checkpoint API

Pause, checkpoint, and resume debates.

List Checkpoints

GET /api/checkpoints?debate_id=abc123&limit=20

Response:

{
  "checkpoints": [
    {
      "id": "cp-123",
      "debate_id": "debate-456",
      "round": 3,
      "created_at": "2026-01-18T10:00:00Z",
      "reason": "manual_pause",
      "resumable": true
    }
  ]
}

Get Resumable Checkpoints

GET /api/checkpoints/resumable

Returns checkpoints that can be resumed.

Get Checkpoint Details

GET /api/checkpoints/\{id\}

Response:

{
  "id": "cp-123",
  "debate_id": "debate-456",
  "round": 3,
  "state": {
    "current_round": 3,
    "agents": ["claude-3", "gpt-4"],
    "votes": {"claude-3": 2, "gpt-4": 1}
  },
  "created_at": "2026-01-18T10:00:00Z",
  "resumable": true
}

Resume from Checkpoint

POST /api/checkpoints/\{id\}/resume
Authorization: Bearer <token>

Resumes a paused debate from the checkpoint state.

Delete Checkpoint

DELETE /api/checkpoints/\{id\}
Authorization: Bearer <token>

Create Checkpoint for Debate

POST /api/debates/\{id\}/checkpoint
Authorization: Bearer <token>
Content-Type: application/json

{
  "reason": "user_requested"
}

Pause Debate

POST /api/debates/\{id\}/pause
Authorization: Bearer <token>

Pauses a running debate and creates a checkpoint automatically.

---

Workspace API (Enterprise)

Multi-tenant workspace management with data isolation.

Create Workspace

POST /api/workspaces
Authorization: Bearer <token>
Content-Type: application/json

{
  "name": "Engineering Team",
  "description": "Workspace for engineering debates",
  "settings": {
    "default_retention_days": 90,
    "require_approval": true
  }
}

List Workspaces

GET /api/workspaces
Authorization: Bearer <token>

Get Workspace Details

GET /api/workspaces/\{id\}
Authorization: Bearer <token>

Delete Workspace

DELETE /api/workspaces/\{id\}
Authorization: Bearer <token>

Add Workspace Member

POST /api/workspaces/\{id\}/members
Authorization: Bearer <token>
Content-Type: application/json

{
  "user_id": "user-123",
  "role": "member",
  "permissions": ["read", "write"]
}

Remove Workspace Member

DELETE /api/workspaces/\{id\}/members/\{user_id\}
Authorization: Bearer <token>

Retention Policies

GET /api/retention/policies
POST /api/retention/policies
PUT /api/retention/policies/\{id\}
DELETE /api/retention/policies/\{id\}
POST /api/retention/policies/\{id\}/execute
GET /api/retention/expiring

Content Classification

POST /api/classify
Content-Type: application/json

{
  "content": "Debate transcript text...",
  "context": "internal_review"
}

Response:

{
  "level": "confidential",
  "confidence": 0.92,
  "reasons": ["contains_internal_data"]
}

Audit Log

GET /api/audit/entries?actor=user-123&action=read&limit=100
GET /api/audit/report?start=2026-01-01&end=2026-01-18
GET /api/audit/verify

SOC 2 Controls: CC6.1, CC6.3 - Logical access controls

---

Privacy API (GDPR/CCPA)

Data export and account deletion for regulatory compliance.

Export User Data

GET /api/v1/privacy/export
Authorization: Bearer <token>

GDPR Article 15, CCPA Right to Know. Exports all user data.

Response:

{
  "export_id": "export-123",
  "status": "processing",
  "estimated_completion": "2026-01-18T10:30:00Z"
}

Get Data Inventory

GET /api/v1/privacy/data-inventory
Authorization: Bearer <token>

Response:

{
  "categories": [
    {
      "name": "debate_history",
      "description": "Your debate participation history",
      "count": 42,
      "retention": "90 days"
    },
    {
      "name": "preferences",
      "description": "Account settings and preferences",
      "count": 1,
      "retention": "Until account deletion"
    }
  ]
}

Delete Account

DELETE /api/v1/privacy/account
Authorization: Bearer <token>
Content-Type: application/json

{
  "password": "your-current-password",
  "confirm": true,
  "reason": "no_longer_needed"
}

GDPR Article 17, CCPA Right to Delete.

Update Privacy Preferences

POST /api/v1/privacy/preferences
Authorization: Bearer <token>
Content-Type: application/json

{
  "do_not_sell": true,
  "marketing_opt_out": false,
  "analytics_opt_out": false,
  "third_party_sharing": true
}

CCPA Do Not Sell compliance.

SOC 2 Control: P5-01 - User access to personal data

---

Training Export API

Export debate data for model fine-tuning.

Export SFT Data

POST /api/training/export/sft
Authorization: Bearer <token>
Content-Type: application/json

{
  "filters": {
    "min_quality": 0.8,
    "start_date": "2026-01-01",
    "end_date": "2026-01-18"
  },
  "format": "jsonl"
}

Supervised Fine-Tuning format export.

Export DPO Data

POST /api/training/export/dpo
Authorization: Bearer <token>
Content-Type: application/json

{
  "filters": {
    "min_margin": 0.3,
    "include_ties": false
  },
  "format": "jsonl"
}

Direct Preference Optimization format with chosen/rejected pairs.

Export Gauntlet Data

POST /api/training/export/gauntlet
Authorization: Bearer <token>

Adversarial training data from gauntlet challenges.

Get Export Statistics

GET /api/training/stats

Response:

{
  "total_debates": 1234,
  "exportable_sft": 890,
  "exportable_dpo": 456,
  "last_export": "2026-01-17T15:00:00Z"
}

List Export Formats

GET /api/training/formats

Response:

{
  "formats": [
    {"id": "jsonl", "name": "JSON Lines", "extensions": [".jsonl"]},
    {"id": "parquet", "name": "Apache Parquet", "extensions": [".parquet"]},
    {"id": "csv", "name": "CSV", "extensions": [".csv"]}
  ]
}

Training Jobs

GET /api/training/jobs
POST /api/training/jobs/\{id\}/start
POST /api/training/jobs/\{id\}/complete
GET /api/training/jobs/\{id\}/metrics
GET /api/training/jobs/\{id\}/artifacts

---

Explainability API

Get decision explanations, evidence chains, vote pivots, counterfactuals, and summaries.

Get Full Explanation

GET /api/v1/debates/\{debate_id\}/explanation
Authorization: Bearer <token>

Query Parameters:

• format (string, default: json): json or summary (Markdown)

Response (json):

{
  "decision_id": "dec-abc123",
  "debate_id": "debate-123",
  "conclusion": "Adopt a tiered rate limiting strategy",
  "confidence": 0.87,
  "consensus_reached": true,
  "evidence_chain": [],
  "vote_pivots": [],
  "counterfactuals": [],
  "evidence_quality_score": 0.82
}

Get Evidence Chain

GET /api/v1/debates/\{debate_id\}/evidence
Authorization: Bearer <token>

Query Parameters:

• limit (int, default: 20)
• min_relevance (float, default: 0.0)

Response:

{
  "debate_id": "debate-123",
  "evidence_count": 5,
  "evidence_quality_score": 0.82,
  "evidence": []
}

Get Vote Pivots

GET /api/v1/debates/\{debate_id\}/votes/pivots
Authorization: Bearer <token>

Query Parameters:

• min_influence (float, default: 0.0)

Get Counterfactuals

GET /api/v1/debates/\{debate_id\}/counterfactuals
Authorization: Bearer <token>

Query Parameters:

• min_sensitivity (float, default: 0.0)

Get Summary

GET /api/v1/debates/\{debate_id\}/summary
Authorization: Bearer <token>

Query Parameters:

• format (string, default: markdown): markdown, json, or html

Batch Explainability

Process multiple debates in a single request for efficiency.

Create Batch Job

POST /api/v1/explainability/batch
Authorization: Bearer <token>
Content-Type: application/json

{
  "debate_ids": ["debate-1", "debate-2", "debate-3"],
  "options": {
    "include_evidence": true,
    "include_counterfactuals": false,
    "include_vote_pivots": false,
    "format": "full"  // "full", "summary", or "minimal"
  }
}

Response (202 Accepted):

{
  "batch_id": "batch-a1b2c3d4e5f6",
  "status": "pending",
  "total_debates": 3,
  "status_url": "/api/v1/explainability/batch/batch-a1b2c3d4e5f6/status",
  "results_url": "/api/v1/explainability/batch/batch-a1b2c3d4e5f6/results"
}

Get Batch Status

GET /api/v1/explainability/batch/\{batch_id\}/status
Authorization: Bearer <token>

Response:

{
  "batch_id": "batch-a1b2c3d4e5f6",
  "status": "processing",
  "total_debates": 3,
  "processed_count": 2,
  "success_count": 2,
  "error_count": 0,
  "created_at": 1737356400.0,
  "started_at": 1737356401.0,
  "progress_pct": 66.7
}

Get Batch Results

GET /api/v1/explainability/batch/\{batch_id\}/results
Authorization: Bearer <token>

Query Parameters:

• include_partial (boolean, default: false): Include results while still processing
• offset (int, default: 0): Pagination offset
• limit (int, default: 50, max: 100): Results per page

Response:

{
  "batch_id": "batch-a1b2c3d4e5f6",
  "status": "completed",
  "total_debates": 3,
  "processed_count": 3,
  "success_count": 3,
  "error_count": 0,
  "results": [
    {
      "debate_id": "debate-1",
      "status": "success",
      "processing_time_ms": 125.4,
      "explanation": {
        "confidence": 0.87,
        "consensus_reached": true,
        "contributing_factors": [...]
      }
    }
  ],
  "pagination": {
    "offset": 0,
    "limit": 50,
    "total": 3,
    "has_more": false
  }
}

Compare Explanations

Compare decision factors across multiple debates.

POST /api/v1/explainability/compare
Authorization: Bearer <token>
Content-Type: application/json

{
  "debate_ids": ["debate-1", "debate-2"],
  "compare_fields": ["confidence", "consensus_reached", "contributing_factors", "evidence_quality"]
}

Response:

{
  "debates_compared": ["debate-1", "debate-2"],
  "comparison": {
    "confidence": {
      "debate-1": 0.87,
      "debate-2": 0.92
    },
    "confidence_stats": {
      "min": 0.87,
      "max": 0.92,
      "avg": 0.895,
      "spread": 0.05
    },
    "consensus_reached": {
      "debate-1": true,
      "debate-2": true
    },
    "consensus_agreement": true,
    "contributing_factors": {
      "evidence_quality": {
        "debate-1": 0.35,
        "debate-2": 0.42
      }
    },
    "common_factors": ["evidence_quality", "argument_strength"]
  }
}

---

Workflow Templates API

Manage and execute reusable workflow templates with pre-built patterns.

List Workflow Templates

GET /api/workflow/templates
Authorization: Bearer <token>

Query Parameters:

• category (string): Filter by category
• pattern (string): Filter by pattern type (hive_mind, map_reduce, review_cycle, pipeline, parallel)
• search (string): Search by name or description
• tags (string): Filter by tags (comma-separated)
• limit (int, default: 50, max: 100)
• offset (int, default: 0)

Response:

{
  "templates": [
    {
      "id": "security/code-review",
      "name": "Security Code Review",
      "description": "Multi-agent security analysis workflow",
      "category": "security",
      "pattern": "review_cycle",
      "tags": ["security", "code", "audit"]
    }
  ],
  "total": 15,
  "limit": 50,
  "offset": 0
}

Get Template Details

GET /api/workflow/templates/\{template_id\}
Authorization: Bearer <token>

Get Template Package

GET /api/workflow/templates/\{template_id\}/package
Authorization: Bearer <token>

Query Parameters:

• include_examples (boolean, default: true): Include usage examples

Response:

{
  "id": "security/code-review",
  "name": "Security Code Review",
  "description": "Multi-agent security analysis workflow",
  "workflow_definition": {
    "nodes": [...],
    "edges": [...],
    "config": {...}
  },
  "input_schema": {
    "type": "object",
    "properties": {
      "code_path": {"type": "string"},
      "severity_threshold": {"type": "string", "enum": ["low", "medium", "high"]}
    }
  },
  "output_schema": {...},
  "documentation": "# Security Code Review\n\nThis workflow...",
  "examples": [
    {
      "name": "Basic usage",
      "inputs": {"code_path": "src/auth.py"},
      "description": "Review authentication module"
    }
  ],
  "version": "1.2.0"
}

Run Workflow Template

POST /api/workflow/templates/\{template_id\}/run
Authorization: Bearer <token>
Content-Type: application/json

{
  "inputs": {
    "code_path": "src/api/handlers.py",
    "severity_threshold": "medium"
  },
  "config": {
    "timeout": 300,
    "priority": "high",
    "async": false
  },
  "workspace_id": "ws-123"
}

Response (sync):

{
  "execution_id": "exec-456",
  "status": "completed",
  "result": {
    "findings": [...],
    "summary": "Found 3 potential vulnerabilities"
  },
  "duration_ms": 45200
}

Response (async):

{
  "execution_id": "exec-456",
  "status": "running",
  "status_url": "/api/workflow-executions/exec-456"
}

List Template Categories

GET /api/workflow/categories

Response:

{
  "categories": [
    {
      "id": "security",
      "name": "Security",
      "description": "Security analysis workflows",
      "template_count": 8,
      "icon": "shield"
    },
    {
      "id": "research",
      "name": "Research",
      "description": "Research and analysis workflows",
      "template_count": 12,
      "icon": "search"
    }
  ]
}

List Workflow Patterns

GET /api/workflow/patterns

Response:

{
  "patterns": [
    {
      "id": "hive_mind",
      "name": "Hive Mind",
      "description": "Parallel agent consultation with aggregation",
      "available": true,
      "use_cases": ["Research", "Brainstorming", "Consensus building"]
    },
    {
      "id": "map_reduce",
      "name": "Map-Reduce",
      "description": "Distribute work, combine results",
      "available": true,
      "use_cases": ["Large document analysis", "Batch processing"]
    }
  ]
}

Instantiate Pattern

POST /api/workflow/patterns/\{pattern_id\}/instantiate
Authorization: Bearer <token>
Content-Type: application/json

{
  "name": "My Security Review",
  "description": "Custom security review workflow",
  "category": "security",
  "config": {
    "max_agents": 5,
    "consensus_threshold": 0.7
  },
  "agents": ["anthropic-api", "openai-api", "mistral-api"]
}

---

Gauntlet Receipts API

Access and export gauntlet verification receipts with cryptographic integrity verification.

List Gauntlet Receipts

GET /api/gauntlet/receipts
Authorization: Bearer <token>

Query Parameters:

• verdict (string): Filter by verdict (pass, fail, warn)
• from_date (string): Start date (ISO 8601)
• to_date (string): End date (ISO 8601)
• limit (int, default: 50)
• offset (int, default: 0)

Response:

{
  "receipts": [
    {
      "id": "receipt-789",
      "debate_id": "debate-123",
      "verdict": "pass",
      "confidence": 0.92,
      "timestamp": "2026-01-20T10:30:00Z",
      "hash": "sha256:a1b2c3..."
    }
  ],
  "total": 42,
  "limit": 50,
  "offset": 0
}

Get Receipt Details

GET /api/gauntlet/receipts/\{receipt_id\}
Authorization: Bearer <token>

Response:

{
  "id": "receipt-789",
  "debate_id": "debate-123",
  "verdict": "pass",
  "confidence": 0.92,
  "timestamp": "2026-01-20T10:30:00Z",
  "hash": "sha256:a1b2c3d4e5f6...",
  "agents": ["anthropic-api", "openai-api"],
  "rounds": 3,
  "consensus_type": "majority",
  "findings": [
    {
      "type": "security",
      "severity": "low",
      "description": "Input validation could be strengthened"
    }
  ],
  "metadata": {
    "duration_ms": 45000,
    "token_count": 12500
  }
}

Verify Receipt Integrity

GET /api/gauntlet/receipts/\{receipt_id\}/verify
Authorization: Bearer <token>

Response:

{
  "valid": true,
  "hash": "sha256:a1b2c3d4e5f6...",
  "computed_hash": "sha256:a1b2c3d4e5f6...",
  "verification_timestamp": "2026-01-20T11:00:00Z"
}

Export Receipt

GET /api/gauntlet/receipts/\{receipt_id\}/export
Authorization: Bearer <token>

Query Parameters:

• format (string, required): "json", "html", "markdown", or "sarif"

Response (format=sarif):

{
  "$schema": "https://raw.githubusercontent.com/oasis-tcs/sarif-spec/master/Schemata/sarif-schema-2.1.0.json",
  "version": "2.1.0",
  "runs": [
    {
      "tool": {
        "driver": {
          "name": "Aragora Gauntlet",
          "version": "1.0.0"
        }
      },
      "results": [...]
    }
  ]
}

Create Shareable Link

POST /api/gauntlet/receipts/\{receipt_id\}/share
Authorization: Bearer <token>
Content-Type: application/json

{
  "expires_in": 86400,
  "allow_download": true
}

Response:

{
  "share_url": "https://app.aragora.ai/receipts/share/abc123",
  "expires_at": "2026-01-21T11:00:00Z",
  "token": "abc123"
}

Get Receipt Statistics

GET /api/gauntlet/receipts/stats
Authorization: Bearer <token>

Query Parameters:

• period (string): "day", "week", "month"

Response:

{
  "period": "week",
  "total": 156,
  "by_verdict": {
    "pass": 120,
    "fail": 24,
    "warn": 12
  },
  "average_confidence": 0.87,
  "average_duration_ms": 42000
}

---

Deprecated Endpoints

The following endpoints are deprecated and will be removed in future versions.

Endpoint	Status	Removal	Migration
GET /api/debates/list	Deprecated	v2.0	Use GET /api/debates
POST /api/debate/new	Deprecated	v2.0	Use POST /api/debates/start
GET /api/elo/rankings	Deprecated	v2.0	Use GET /api/agent/leaderboard
GET /api/agent/elo	Deprecated	v2.0	Use GET /api/agent/\{name\}/profile
POST /api/stream/start	Deprecated	v2.0	Use WebSocket /ws connection

Migration Guide

Debate Listing

# Old (deprecated)
curl http://localhost:8080/api/debates/list

# New (recommended)
curl http://localhost:8080/api/debates

Starting Debates

# Old (deprecated)
curl -X POST http://localhost:8080/api/debate/new \
  -d '{"question": "..."}'

# New (recommended)
curl -X POST http://localhost:8080/api/debates/start \
  -H "Content-Type: application/json" \
  -d '{"question": "...", "agents": ["anthropic-api", "openai-api"]}'

Agent Rankings

# Old (deprecated)
curl http://localhost:8080/api/elo/rankings
curl http://localhost:8080/api/agent/elo?name=anthropic-api

# New (recommended)
curl http://localhost:8080/api/agent/leaderboard
curl http://localhost:8080/api/agent/anthropic-api/profile

Deprecation Timeline

• v1.5 (Current): Deprecated endpoints still functional, emit deprecation warnings in response headers
• v2.0 (Planned): Deprecated endpoints return 410 Gone status
• v2.1 (Planned): Deprecated endpoints removed entirely

Response headers for deprecated endpoints:

Deprecation: true
Sunset: 2026-06-01
Link: </api/debates>; rel="successor-version"

---

Changelog

2026-01-09

• Added Graph Debates API (4 endpoints for branching debates)
• Added Matrix Debates API (4 endpoints for parallel scenarios)
• Added Breakpoints API (3 endpoints for human-in-the-loop)
• Added Introspection API (4 endpoints for agent self-awareness)
• Added Gallery API (3 endpoints for public debate sharing)
• Added Billing API (8 endpoints for subscription management)
• Updated endpoint count from 106 to 124

2026-01-05

• Added CSV and HTML export formats for debates
• Added token streaming events (token_start, token_delta, token_end)
• Added phase timeout events (phase_start now includes timeout, added phase_timeout)
• Added usage examples section with curl commands
• Added deprecated endpoints documentation with migration guide