Assistix products
Assistix LLM Gateway
OpenAI-compatible API for Assistix products. Externally you call Tide / Flare / Pine. Assistix maps them to upstream providers; the provider key never leaves the server.
Base URLhttps://api.assistix.org/v1
AuthAuthorization: Bearer <assistix_id_access_token>
Optional: X-Assistix-Product: beam-agent
Models
| id | Name | Role |
|---|---|---|
tide | Tide | Fast / everyday |
flare | Flare | Balanced / long context |
pine | Pine | Strong reasoning |
Raw upstream model ids are rejected. GET /v1/models returns only these three.
Chat completions
POST /v1/chat/completions
Content-Type: application/json
Authorization: Bearer <assistix_id_access_token>
{
"model": "tide",
"messages": [{"role":"user","content":"Hello"}],
"stream": false
}
Request body is OpenAI-compatible (messages, tools, tool_choice, temperature, max_tokens, stream, …). Assistix rewrites only model before upstream and rewrites it back to tide|flare|pine in the response.
Streaming
Set "stream": true. Response is SSE (text/event-stream), same shape as OpenAI / KumoRouter.
Endpoints
GET /health— liveness (no secrets)GET /v1/models— Assistix catalogPOST /v1/chat/completions— generate
Errors
401— missing/invalid Assistix ID token400— unknownmodel(not tide/flare/pine)429— weekly quota exceeded5xx— upstream / gateway