> ## Documentation Index > Fetch the complete documentation index at: https://docs.ryvo.so/llms.txt > Use this file to discover all available pages before exploring further. # Publish an agent > Engine only. Moves the published pointer to the head version, so the latest saved configuration starts answering. Publishing wakes a `draft` agent (it becomes `active`) but leaves a `paused` one paused. The agent needs a `system_prompt`, an `llm` block and, for voice, an `stt` block and a `voice` with `voice_id`. ## OpenAPI ````yaml /openapi.yaml post /v1/agents/{id}/publish openapi: 3.1.0 info: title: Ryvo API version: 1.0.0 description: > The Ryvo public API lets you drive your AI voice and chat agents from your own systems: trigger calls, read conversations and transcripts, manage agents, knowledge base documents, phone numbers, outbound campaigns and webhooks, and read billing, team and log data. It is the same platform the portal uses, exposed over plain HTTPS and JSON. **Base URL**: `https://api.ryvo.so` ## Authentication Every request (except `GET /v1/models`) carries an API key as a Bearer token: ``` Authorization: Bearer ryvo_live_... ``` Keys are created in the portal under **Settings > API keys** (`https://app.ryvo.so/developers`). The full token is shown once at creation time. The API is available from the Starter plan up. ## Scopes Each key carries a list of scopes in the form `resource:action`, chosen when the key is created. An endpoint requires exactly one scope (listed on each operation as `x-scope`). A key without it receives `403 insufficient_scope`, and the message names the missing scope. | Scope | Opens | |---|---| | `agents:read`, `agents:write` | `/v1/agents` | | `calls:read`, `calls:write` | `/v1/calls` | | `conversations:read` | `/v1/conversations` | | `knowledge:read`, `knowledge:write` | `/v1/knowledge` | | `phone_numbers:read`, `phone_numbers:write` | `/v1/phone-numbers` | | `campaigns:read`, `campaigns:write` | `/v1/batch-calls` | | `webhooks:read`, `webhooks:write` | `/v1/webhooks` | | `integrations:read` | `/v1/integrations` | | `team:read` | `/v1/team` | | `billing:read` | `/v1/billing` | | `logs:read` | `/v1/logs` | `GET /v1/me` only needs a valid key and reports the scopes it carries. Integrations, team and billing are read-only by design: connecting an app is a browser OAuth flow, inviting people is persistent access, and buying credits from a loop is the failure mode nobody wants. ## Pagination Every list takes `?limit=` (1 to 100, default 20) and `?cursor=`, and answers `{ "data": [...], "has_more": true|false, "next_cursor": "..."|null }`. The cursor is opaque: pass `next_cursor` back as `?cursor=` to fetch the following page, and stop when `has_more` is `false`. ## Idempotency `POST /v1/calls` accepts an `Idempotency-Key` header (1 to 255 printable ASCII characters). Retrying the same key with the same body returns the original response for 24 hours without dialing again; the same key with a different body returns `409 idempotency_conflict`. ## Rate limits Two limits apply, in this order, on fixed one-minute windows: 1. **Per IP address**: 300 requests/min across every `/v1` endpoint, except `POST /v1/calls`, which has its own bucket of 10 requests/min per IP because it dials a real phone. 2. **Per account, by plan**: shared by every key of the account (Starter 10, Pro 40, Scale and Enterprise 120 requests/min). A `429 rate_limit_exceeded` response carries a `Retry-After` header with the seconds to wait (always at least 1). ## Errors and request ids Every error uses the same envelope: `{ "error", "message", "request_id" }`. Branch on `error`, which is a stable code; `message` is for humans and may change. The stable codes are `unauthorized`, `forbidden`, `insufficient_scope`, `rate_limit_exceeded`, `invalid_json`, `invalid_body`, `invalid_query`, `not_found`, `conflict`, `unsupported_runtime`, `plan_limit`, `payment_required`, `upstream_error` and `internal_error`. `POST /v1/calls` keeps a few older codes of its own (`agent_not_found`, `agent_disabled`, `idempotency_conflict`, `invalid_idempotency_key`, `key_budget_exhausted`, `upstream_unreachable`, `service_unavailable`). Every response, success or error, carries an `X-Request-Id` header. Quote it when you report a problem. A `404 not_found` is returned both when a resource does not exist and when it belongs to another account, with the same message: the API does not confirm which ids exist. contact: name: Ryvo support email: team@ryvo.so url: https://ryvo.so servers: - url: https://api.ryvo.so description: Production security: - bearerAuth: [] tags: - name: Account description: Who the key belongs to and what it can do. - name: Agents description: >- Your agents on both runtimes. Engine agents are fully manageable (configuration, versions, publish). ElevenLabs agents are listed and can be renamed; their configuration lives in the portal. - name: Calls description: >- Voice conversations (phone and web voice) on both runtimes, and the endpoint to trigger an outbound call. - name: Conversations description: >- Every conversation on every channel (voice, chat, WhatsApp), with the same shape as calls plus the message count. - name: Knowledge base description: >- Documents and folders that Engine agents consult during a conversation. Requires the Engine module on your plan. - name: Phone numbers description: Numbers imported from your Twilio account and their agent assignment. - name: Batch calls description: >- Outbound campaigns on the ElevenLabs runtime. Uses the `campaigns:*` scopes. - name: Webhooks description: Outbound endpoints, their deliveries and the event catalog. - name: Integrations description: >- The integration catalog with the connection status of your account. Read-only. - name: Team description: Members, invitations and seats. Read-only. - name: Billing description: Plan, credits, wallet ledger, invoices and daily usage. Read-only. - name: Logs description: Requests made to this API and the account activity log. - name: Models description: The public model catalog with credits and USD per plan. No authentication. paths: /v1/agents/{id}/publish: parameters: - $ref: '#/components/parameters/agentId' post: tags: - Agents summary: Publish an agent description: | Engine only. Moves the published pointer to the head version, so the latest saved configuration starts answering. Publishing wakes a `draft` agent (it becomes `active`) but leaves a `paused` one paused. The agent needs a `system_prompt`, an `llm` block and, for voice, an `stt` block and a `voice` with `voice_id`. operationId: publishAgent responses: '200': description: The version now serving traffic and the agent status. content: application/json: schema: type: object required: - id - published_version - status properties: id: type: string format: uuid published_version: type: integer status: $ref: '#/components/schemas/AgentStatus' example: id: 4d0f3c2a-6b7e-4a1c-9f2d-1e2b3c4d5e6f published_version: 3 status: active '401': $ref: '#/components/responses/Unauthorized' '402': description: >- The account needs a verified email, a verified phone and a payment method on file before publishing (`payment_required`). content: application/json: schema: $ref: '#/components/schemas/Error' '403': description: Missing scope, or the Engine module is not enabled (`plan_limit`). content: application/json: schema: $ref: '#/components/schemas/Error' '404': $ref: '#/components/responses/NotFound' '409': description: >- `unsupported_runtime` for ElevenLabs agents; `conflict` when the agent is archived, has no saved configuration or is missing a required block. content: application/json: schema: $ref: '#/components/schemas/Error' '429': $ref: '#/components/responses/RateLimited' components: parameters: agentId: name: id in: path required: true description: The Ryvo uuid of the agent. schema: type: string format: uuid example: 4d0f3c2a-6b7e-4a1c-9f2d-1e2b3c4d5e6f schemas: AgentStatus: type: string enum: - draft - active - paused - archived - error example: active Error: type: object required: - error - message - request_id properties: error: type: string description: A stable error code. Branch on this, not on `message`. example: unauthorized message: type: string description: A human-readable description. May change between versions. example: Invalid or revoked API key. request_id: type: string format: uuid description: The request id. Quote it when you report a problem. example: f0c0a9d2-8b1e-4d8a-9c2f-6e5b7a1d2c3b responses: Unauthorized: description: >- Missing or malformed `Authorization` header, or an invalid, expired or revoked key (`unauthorized`). content: application/json: schema: $ref: '#/components/schemas/Error' example: error: unauthorized message: Invalid or revoked API key. request_id: f0c0a9d2-8b1e-4d8a-9c2f-6e5b7a1d2c3b NotFound: description: >- The resource does not exist or belongs to another account (`not_found`). Both cases return the same message. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: not_found message: Resource not found. request_id: f0c0a9d2-8b1e-4d8a-9c2f-6e5b7a1d2c3b RateLimited: description: >- Too many requests for this IP or this account (`rate_limit_exceeded`). Wait `Retry-After` seconds. headers: Retry-After: description: Seconds to wait before retrying. Always at least 1. schema: type: integer example: 32 content: application/json: schema: $ref: '#/components/schemas/Error' example: error: rate_limit_exceeded message: Rate limit exceeded. Retry in 32s (see the Retry-After header). request_id: f0c0a9d2-8b1e-4d8a-9c2f-6e5b7a1d2c3b securitySchemes: bearerAuth: type: http scheme: bearer bearerFormat: API key (ryvo_live_...) description: | A Bearer token. Create keys in the portal under Settings > API keys (`https://app.ryvo.so/developers`). The token has the form `ryvo_live_<64 hex characters>` and is shown once. **Header:** `Authorization: Bearer ryvo_live_...` ````