API overview
Copy Page as Markdown
SubQ serves two model families, each on its own endpoints:
- SubQ Snap (
subq-snap-1.2) returns typed decisions about text and images on the Decisions API. Available to every account. - SubQ Savant (
subq-savant-1.4) Early access generates text over up to 4M tokens of context on Chat Completions and Responses, both OpenAI-compatible. Request access.
The Files, Conversations, and Models endpoints support them. See Models for what each model accepts.
Request flow
- Create an account at console.subq.ai
- Generate an API key from the console
- Set the SubQ base URL in your HTTP client or OpenAI SDK
- Choose an endpoint and send a request using a SubQ model.
- Enable streaming on Chat Completions or Responses when your product needs tokens as they are generated
Base URL
Send API requests to:
https://api.subq.ai/v1Append the endpoint path to it: Chat Completions is
https://api.subq.ai/v1/chat/completions. In an OpenAI SDK, set baseURL
(TypeScript) or base_url (Python) to the base URL and the SDK adds the path.
Treat each endpoint page as the contract for its supported models, fields, and response formats.
Endpoints
| Endpoint | Method | Path | Model | Use it to |
|---|---|---|---|---|
| Models | GET | /v1/models | — | List the model IDs available to your API key at runtime. |
| Decisions | POST | /v1/systemone | subq-snap-1.2 | Return typed yes/no, choice, and score decisions about text or images, with probabilities and confidence. |
| Chat Completions | POST | /v1/chat/completions | subq-savant-1.4 Early access | Generate text from messages, statelessly. |
| Responses | POST | /v1/responses | subq-savant-1.4 Early access | Run multi-turn and long-context workflows with instructions, built-in web search, and server-side history via previous_response_id. |
| Conversations | POST / GET / DELETE | /v1/conversations | Used with subq-savant-1.4 | Create, inspect, retag, and delete a conversation's history, independent of any single turn. |
| Files | POST / GET / DELETE | /v1/files | Used with subq-savant-1.4 | Upload documents once and reference them by file_id in Chat Completions or Responses. |
Next steps
- Use the Quickstart to verify your API key.
- Read Authentication before wiring credentials into an app.
- Review capabilities, context windows, and pricing in Models.
- Return typed decisions with the Decisions API.
- Generate text with the Chat Completions API.
- Put documents in a prompt with Sending files.
- Return schema-constrained JSON with Structured outputs.
- Connect models to your app with Function calling.
- Check Data retention for how long SubQ stores your prompts and completions.