API Reference

BeLive Studio API overview — base URL, authentication, error codes and compatibility.

BeLive Studio is a unified LLM API gateway. All endpoints follow the OpenAI-compatible format, with native-format compatibility for Claude and Gemini.

Full API docs

This section covers the core endpoints. Use the current console endpoint definitions for the complete list, field details, and online debugging.

See Limits and Quotas for gateway request-body and multimodal file limits.

Base URL

https://api.tokenhub.belivestudios.com

Authentication

All requests carry the API key in an HTTP header:

Authorization: Bearer sk-your-belive-studios-key

API keys are created in "API Tokens" on the BeLive Studio Console, with optional model, IP and quota restrictions.

Quick Test

curl https://api.tokenhub.belivestudios.com/v1/chat/completions \
  -H "Authorization: Bearer sk-your-belive-studios-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-sol",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Endpoints at a Glance

Category	Endpoint	Description
Chat	`POST /v1/chat/completions`	OpenAI-compatible chat with streaming, tool calls and multimodality
Chat	`POST /v1/messages`	Claude (Anthropic) native Messages endpoint
Chat	`POST /v1beta/models/{model}:generateContent`	Gemini (Google) native non-streaming generation
Chat	`POST /v1beta/models/{model}:streamGenerateContent`	Gemini native streaming generation
Image	`POST /v1/images/generations`	GPT Image text-to-image, returning Base64 image data
Image	`POST /v1/images/edits`	Edit or extend images with GPT Image and a reference image
Image	`POST /v1/images/generations`	GPT Image common endpoint: generate or edit depending on `image`
Image	`POST /v1/images/generations`	Seedream text-to-image and image-to-image
Image	`POST /v1/models/{model}:generateContent`	Native Gemini image generation and editing
Image	`POST /v1/images/generations`	OpenAI-compatible Gemini image endpoint, returning an image URL
Video	`POST /api/v3/contents/generations/tasks`	Create a Seedance video generation task
Video	`GET /api/v3/contents/generations/tasks/{task_id}`	Query a Seedance task status and result
Video	MiniMax H3	Create and query MiniMax H3 / H3 Max video generation tasks
Video	Kling 3.0 Turbo	Text-to-video: `POST /text-to-video/kling-3.0-turbo`; image-to-video: `POST /image-to-video/kling-3.0-turbo`
Video	Kling 3.0 / 3.0 Omni	Text: `POST /text-to-video/kling-3.0`; image: `POST /image-to-video/kling-3.0`; Omni: `POST /omni-video/kling-3.0-omni`; motion control: `POST /motion-control/kling-3.0`
Video	Kling O1	`POST /omni-video/kling-o1` for text, image and video inputs
Video	`GET /tasks`	Query Kling asynchronous tasks by task ID
Video	`POST /tasks`	Paginate Kling asynchronous tasks by time, status or product type
Management	`GET /v1/models`	List models callable by the current token
Management	`GET /v1/dashboard/billing/subscription`	Query token balance and usage

Error Codes

HTTP status	Meaning
400	Bad request (misspelled model, missing params, etc.)
401	Invalid or missing API key
403	Token lacks permission (model not authorized, IP restricted, quota exhausted)
404	Wrong path (check the `/v1` prefix)
429	Rate limited
500 / 502 / 503	Upstream or gateway error — retryable

Error responses are JSON with error.message, error.type and error.code.