API
OpenAPI specification (Swagger)
OpenAPI 3.1 specification of the SovrGPT API: machine-readable, importable into Swagger UI, Redoc, Postman, Insomnia and client SDK code generators.
The SovrGPT API is available as an OpenAPI 3.1 specification. Import it into common tools or generate client SDKs from it.
Specification: https://sovrgpt.com/openapi.json
Using it
- Swagger UI / Swagger Editor: open editor.swagger.io →
File → Import URL →
https://sovrgpt.com/openapi.json. - Redoc:
npx @redocly/cli preview-docs https://sovrgpt.com/openapi.json. - Postman / Insomnia: Import → Link → paste the URL. You get a collection with every endpoint.
- Generating client SDKs:
openapi-generator,openapi-typescriptand others read the file directly.
Endpoints covered
Paths are relative to the base URL https://sovrgpt.com/api/v1.
| Endpoint | Purpose | Scope |
|---|---|---|
GET /models | Query the available models (chat, embeddings, rerank, TTS, STT, decisions) | – |
GET /me | Permissions of the calling key | – |
POST /chat/completions | Chat completion | chat |
POST /embeddings | Text embeddings | embeddings |
POST /rerank | Relevance re-ordering | rerank |
POST /audio/speech | Text-to-speech (voice catalogue, Supertonic, CosyVoice) | speech |
GET /audio/voices, GET /audio/voices/design-options | Voice catalogue, casting sheet | – |
/audio/voices/designs with /{id}, /{id}/adopt, /suggest | Design, review and adopt your own voices | speech |
POST /audio/transcriptions | Speech-to-text (Voxtral) | transcribe |
POST /audio/uploads | Upload ticket for large audio files | transcribe |
POST /decisions | Request a decision | decisions |
GET /decisions/presets, GET /decisions/models | Question packs, decision models | – |
The MCP endpoints (/mcp/mcp, /mcp/docs/mcp) are not in the specification; see
MCP server.
The specification is the machine-readable reference. Explained examples live on the pages for chat, models, audio and decisions.