API Reference
RESTful endpoints for the BOSS platform. Authenticate with an API key, call any endpoint, and receive structured JSON responses. All examples include both curl and TypeScript SDK versions.
Authentication
All API requests require a Bearer token in the Authorization header. Generate API keys at /admin/settings/api-keys. Each key has configurable scopes that limit what endpoints it can access.
sk_cintrico_[prefix]_[random]Available scopes:
studios:readstudios:writedocuments:readdocuments:writeagents:invokeagents:managechat:readchat:writesearch:querywebhooks:manageadmin:readadmin:writecurl example
TypeScript SDK
Endpoints
Request & response examples
POST /api/chat
Send a message to an agent and receive a structured response with optional canvas blocks, citations, and cost tracking.
Response:
POST /api/search
Semantic vector search across all studio content. Returns ranked results with relevance scores, snippets, and metadata.
Response:
GET /api/agents
List all agents with their current trust scores, stats, and configuration.
Response:
Webhook setup
Register outbound webhooks to receive real-time notifications when events occur in your BOSS workspace. All webhook payloads include an HMAC signature for verification.
Available event types:
document.createdNew document created in any studiodocument.publishedDocument moved from draft to publishedtask.completedChecklist item or task marked donetask.overdueTask past its due dateagent.flaggedService Engine flagged an agent actionagent.blockedService Engine blocked an agent actionapproval.requestedAgent requesting human approvalapproval.grantedHuman approved an agent actionstudio.createdNew studio created in workspacemember.joinedNew member added to studioswarm.completedMulti-agent swarm finished executionplaybook.checkpointPlaybook reached a human checkpointError handling
All errors follow a consistent envelope format. The error field contains a machine-readable code and a human-readable message.
Retry strategy with exponential backoff
The recommended pattern for handling transient errors:
Rate limits
Rate limit headers are included in every response:
X-RateLimit-LimitMaximum requests per windowX-RateLimit-RemainingRequests remaining in current windowX-RateLimit-ResetUnix timestamp when the window resetsRetry-AfterSeconds to wait (only on 429 responses)Enterprise plans have higher rate limits (1,000 req/min, 100,000 req/hr) and dedicated infrastructure. Contact sales for custom rate limit configurations.
Response format
All responses follow a consistent envelope format with data, error, and meta fields. Successful responses have error: null. Failed responses have data: null.