# TokenLab

> Unified AI API Gateway - Access hundreds of AI models through a single API

TokenLab is an AI API aggregation platform providing unified access to models from OpenAI, Anthropic, Google, DeepSeek, and more.

## Quick Start

Base URL: https://api.tokenlab.sh/v1
Authentication: Bearer token (API key from dashboard)

```bash
curl https://api.tokenlab.sh/v1/chat/completions \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-4o", "messages": [{"role": "user", "content": "Hello!"}]}'
```

## Documentation

- [Introduction](https://tokenlab.sh/docs/en/introduction): Overview and key features
- [Quickstart](https://tokenlab.sh/docs/en/quickstart): Get started in 2 minutes
- [Authentication](https://tokenlab.sh/docs/en/authentication): API key management
- [Management API](https://tokenlab.sh/docs/en/api-reference/management/introduction): Organization balance, key management, and key-level reporting
- [API Formats](https://tokenlab.sh/docs/en/guides/api-formats): OpenAI, Anthropic, Gemini formats
- [Streaming](https://tokenlab.sh/docs/en/guides/streaming): Real-time responses
- [Error Handling](https://tokenlab.sh/docs/en/guides/error-handling): Error codes and handling
- [Data retention](https://tokenlab.sh/data-retention): Console capture expiry and generated-media retention

## API Reference

- [Chat Completions](https://tokenlab.sh/docs/en/api-reference/chat/create-completion): POST https://api.tokenlab.sh/v1/chat/completions
- [Messages (Anthropic)](https://tokenlab.sh/docs/en/api-reference/messages/create-message): POST https://api.tokenlab.sh/v1/messages
- [Responses (OpenAI)](https://tokenlab.sh/docs/en/api-reference/responses/create-response): POST https://api.tokenlab.sh/v1/responses
- [Gemini](https://tokenlab.sh/docs/en/api-reference/gemini/generate-content): POST https://api.tokenlab.sh/v1beta/models/{model}:generateContent
- [Images](https://tokenlab.sh/docs/en/api-reference/images/create-image): POST https://api.tokenlab.sh/v1/images/generations
- [Video](https://tokenlab.sh/docs/en/api-reference/video/create-video): POST https://api.tokenlab.sh/v1/videos/generations
- [Audio](https://tokenlab.sh/docs/en/api-reference/audio/create-speech): POST https://api.tokenlab.sh/v1/audio/speech
- [Realtime](https://tokenlab.sh/docs/en/api-reference/realtime/connect): WS https://api.tokenlab.sh/v1/realtime?model={model} (WebSocket subset only; no Realtime REST client secrets or Calls)
- [Translations](https://tokenlab.sh/docs/en/api-reference/translation/create-translation): POST https://api.tokenlab.sh/v1/translations
- [Embeddings](https://tokenlab.sh/docs/en/api-reference/embeddings/create-embedding): POST https://api.tokenlab.sh/v1/embeddings
- [Decisions](https://tokenlab.sh/docs/en/api-reference/systemone/create-decision): POST https://api.tokenlab.sh/v1/systemone
- [Models](https://tokenlab.sh/docs/en/api-reference/models/list-models): GET https://api.tokenlab.sh/v1/models
- [Management API](https://tokenlab.sh/docs/en/api-reference/management/introduction): Management token authentication, organization balance, and key-level usage/billing

## Supported Capabilities

| Capability | Example Models |
|------------|----------------|
| Chat | GPT-5.4, Claude Sonnet 4.6, Gemini 3.5 Flash, DeepSeek V4 |
| Vision | GPT-4o, Claude Sonnet 4.6, Gemini 3.5 Flash |
| Image Generation | GPT Image, FLUX, Gemini image models |
| Audio | Whisper, TTS-1, MiniMax Speech 2.8 |
| Realtime WebSocket | GPT realtime and live translation models |
| Embeddings | text-embedding-3-small |
| Music | Suno |
| 3D | Tripo H3.1, Tripo P1.0 |

## Integrations

- [Tool and SDK compatibility](https://tokenlab.sh/docs/en/guides/ide-sdk-compatibility)
- [SDKs & Libraries](https://tokenlab.sh/docs/en/guides/sdks)
- [Control coding agent costs](https://tokenlab.sh/docs/en/guides/coding-agent-cost)
- [Choose a model for coding agents](https://tokenlab.sh/docs/en/guides/coding-agent-models)
- [TokenLab Provider Setup](https://tokenlab.sh/docs/en/guides/tokenlab-provider)
- [TokenLab MCP Server](https://tokenlab.sh/docs/en/integrations/tokenlab-mcp-server)
- [TokenLab Agent Skills](https://tokenlab.sh/docs/en/integrations/coding-agent-skill)
- [DeepSeek Harness](https://tokenlab.sh/docs/en/integrations/deepseek-harness)
- [GitHub Copilot CLI](https://tokenlab.sh/docs/en/integrations/github-copilot-cli)
- [Cursor](https://tokenlab.sh/docs/en/integrations/cursor)
- [Visual Studio Code](https://tokenlab.sh/docs/en/integrations/vscode)
- [Cline](https://tokenlab.sh/docs/en/integrations/cline)
- [Claude Code](https://tokenlab.sh/docs/en/integrations/claude-code)
- [Codex CLI](https://tokenlab.sh/docs/en/integrations/codex-cli)
- [OpenCodex](https://tokenlab.sh/docs/en/integrations/opencodex)
- [WorkBuddy](https://tokenlab.sh/docs/en/integrations/workbuddy)
- [Hermes Agent](https://tokenlab.sh/docs/en/integrations/hermes-agent)
- [Kilo Code](https://tokenlab.sh/docs/en/integrations/kilo-code)
- [Gemini CLI](https://tokenlab.sh/docs/en/integrations/gemini-cli)
- [Pi Coding Agent](https://tokenlab.sh/docs/en/integrations/pi-coding-agent)
- [OpenCode](https://tokenlab.sh/docs/en/integrations/opencode)
- [OpenHands](https://tokenlab.sh/docs/en/integrations/openhands)
- [OpenAI SDK](https://tokenlab.sh/docs/en/integrations/openai-sdk)
- [Anthropic SDK](https://tokenlab.sh/docs/en/integrations/anthropic-sdk)
- [LiteLLM](https://tokenlab.sh/docs/en/integrations/litellm)
- [Vercel AI SDK](https://tokenlab.sh/docs/en/integrations/vercel-ai-sdk)
- [LangChain](https://tokenlab.sh/docs/en/integrations/langchain)
- [LlamaIndex](https://tokenlab.sh/docs/en/integrations/llamaindex)
- [OpenAI Agents SDK](https://tokenlab.sh/docs/en/integrations/openai-agents-sdk)
- [AutoGen](https://tokenlab.sh/docs/en/integrations/autogen)
- [Semantic Kernel](https://tokenlab.sh/docs/en/integrations/semantic-kernel)
- [CrewAI](https://tokenlab.sh/docs/en/integrations/crewai)
- [Pydantic AI](https://tokenlab.sh/docs/en/integrations/pydantic-ai)
- [DSPy](https://tokenlab.sh/docs/en/integrations/dspy)
- [Langflow](https://tokenlab.sh/docs/en/integrations/langflow)
- [Haystack](https://tokenlab.sh/docs/en/integrations/haystack)
- [Graphiti](https://tokenlab.sh/docs/en/integrations/graphiti)
- [Dify](https://tokenlab.sh/docs/en/integrations/dify)
- [Flowise](https://tokenlab.sh/docs/en/integrations/flowise)
- [Mem0](https://tokenlab.sh/docs/en/integrations/mem0)
- [Agno](https://tokenlab.sh/docs/en/integrations/agno)
- [Browser Use](https://tokenlab.sh/docs/en/integrations/browser-use)
- [VoltAgent](https://tokenlab.sh/docs/en/integrations/voltagent)
- [Ragas](https://tokenlab.sh/docs/en/integrations/ragas)
- [Guardrails](https://tokenlab.sh/docs/en/integrations/guardrails)
- [Prompt flow](https://tokenlab.sh/docs/en/integrations/promptflow)
- [Promptfoo](https://tokenlab.sh/docs/en/integrations/promptfoo)
- [Portkey Gateway](https://tokenlab.sh/docs/en/integrations/portkey)
- [Helicone](https://tokenlab.sh/docs/en/integrations/helicone)
- [Langfuse](https://tokenlab.sh/docs/en/integrations/langfuse)
- [OpenLIT](https://tokenlab.sh/docs/en/integrations/openlit)
- [OpenLLMetry](https://tokenlab.sh/docs/en/integrations/openllmetry)
- [LangBot](https://tokenlab.sh/docs/en/integrations/langbot)
- [Phoenix](https://tokenlab.sh/docs/en/integrations/phoenix)
- [Opik](https://tokenlab.sh/docs/en/integrations/opik)
- [AI Chat Applications](https://tokenlab.sh/docs/en/integrations/ai-chat-apps)
- [Open WebUI](https://tokenlab.sh/docs/en/integrations/open-webui)
- [Chatbox](https://tokenlab.sh/docs/en/integrations/chatbox)
- [DeepChat](https://tokenlab.sh/docs/en/integrations/deepchat)
- [Jan](https://tokenlab.sh/docs/en/integrations/jan)
- [LibreChat](https://tokenlab.sh/docs/en/integrations/librechat)
- [Cherry Studio](https://tokenlab.sh/docs/en/integrations/cherry-studio)

## Key Features

- **Native-first + OpenAI-compatible**: Use provider-native routes when behavior matters, or /v1 compatibility for existing OpenAI-style clients
- **Multi-Format API**: Use OpenAI, Anthropic, or Gemini formats with one API key
- **Smart Routing**: Automatic failover and load balancing
- **Pay As You Go**: No subscriptions, competitive pricing
- **Agent-First Model Discovery**: Use GET /v1/models?recommended_for=image|video|music|3d|tts|stt|embedding|rerank|translation|decision before non-chat requests

## Model Data

- [Model Data Center](https://tokenlab.sh/en/models/data): public model catalog source policy and data API
- [Model Trends](https://tokenlab.sh/en/models/trends): newest public models, source coverage, and pricing-unit distribution
- [Model Research](https://tokenlab.sh/en/models/research): model infrastructure research with source snapshots and limitations
- [Catalog JSON](https://tokenlab.sh/model-data/catalog.json): machine-readable public catalog summary
- [Latest Models JSON](https://tokenlab.sh/model-data/latest.json): machine-readable public model list
- [Trends JSON](https://tokenlab.sh/model-data/trends.json): machine-readable current trend snapshot
- [Model Data Markdown](https://tokenlab.sh/model-data/summary.md): crawler-friendly model data summary

## Links

- Website: https://tokenlab.sh
- Dashboard: https://tokenlab.sh/dashboard
- Models: https://tokenlab.sh/en/models
- API Status: https://tokenlab.sh/status
- Support: support@tokenlab.sh