Compatible API guide

OpenAI-Compatible API for Multiple AI Models

Infistar provides a gateway for common OpenAI request formats. Existing OpenAI SDKs, server applications, and clients that accept a custom base URL can usually connect by changing the endpoint and API key, then selecting a model ID from the live catalog.

Use https://infistar.ai/v1 internationally or https://coneverse.com/v1 in mainland China. Authenticate with Authorization: Bearer YOUR_API_KEY. Endpoint support varies by model, so confirm the live model catalog and GET /v1/models response before integrating.

View models and pricingCreate an account

Common OpenAI-compatible endpoints

GET /v1/models

List model IDs available to the current API key

POST /v1/chat/completions

Multi-turn chat and streaming text

POST /v1/responses

Responses API request format

POST /v1/embeddings

Text vectors and embedding models

POST /v1/images/generations

Image models that declare this endpoint

Migrate an existing OpenAI client

  1. Create an API key in the console. Never expose it in browser code, public repositories, or logs.
  2. Set the SDK or client base URL to the /v1 address for the site you are using.
  3. Copy an exact model ID from Models and Pricing or GET /v1/models instead of guessing from a display name.
  4. Test a low-cost non-streaming and streaming request before setting production timeouts, retries, and usage records.

Minimal Chat Completions request

MODEL_ID is a placeholder. Replace it with a currently listed model that explicitly supports Chat Completions.

curl https://infistar.ai/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID",
    "messages": [{"role": "user", "content": "Hello"}],
    "stream": false
  }'

Compatibility boundaries

  • Compatibility covers the common request and response contract published by Infistar; it does not mean every model supports every OpenAI endpoint or optional parameter.
  • Available models depend on the API key group and current public catalog. Do not hard-code a list copied from another account.
  • Pricing dimensions for cache, long context, images, and media vary by model. Use the live pricing page and actual usage record as the final reference.
Anthropic-compatible APIGemini-compatible API