跳到主要内容
Docs/Api Reference/Agents

Agents API

Agents are autonomous goal-driven workers: you set a goal + budget + deadline, and the agent plans and executes a multi-step strategy. Approvals can be required for risky actions.

Endpoints

MethodPathDescription
GET`/api/v1/agents`List with aggregate metrics
POST`/api/v1/agents`Create
GET`/api/v1/agents/:id`Detail (includes recent logs + pending approvals)
POST`/api/v1/agents/:id/start`Start execution
POST`/api/v1/agents/:id/pause`Pause
POST`/api/v1/agents/:id/stop`Stop
GET`/api/v1/agents/:id/log`Execution log (paginated)
DELETE`/api/v1/agents/:id`Delete

Approvals

MethodPathDescription
GET`/api/v1/agents/:id/approvals`Pending approvals
POST`/api/v1/agents/:id/approvals/:approvalId/approve`Approve a step
POST`/api/v1/agents/:id/approvals/:approvalId/reject`Reject

Templates

MethodPathDescription
GET`/api/v1/agents/templates`List templates
POST`/api/v1/agents/templates`Create

Every response uses the standard envelope:

{
  "success": true,
  "data": { /* resource or array */ },
  "meta": { "total": 100, "cursor": "...", "hasMore": true },
  "error": { "code": "ERROR_CODE", "message": "Human readable" }
}

error is omitted on success; meta is omitted when not paginating.

Rate Limits

Subject to the standard 200 req/min/tenant authenticated limit. See [Rate Limits](/docs/api-reference/rate-limits) for full details and the per-endpoint overrides.

← Back to Docs