Skip to main content
The TokenLab MCP server gives MCP-compatible agents a read-only way to discover TokenLab models before they write code or choose an API route. Use it when an agent needs to compare available models, inspect supported request formats, check pricing, or understand which native endpoint family to use for a task.
The public catalog tools do not require a TokenLab API key. The server does not proxy paid inference requests.

​
What It Provides

  • Live model discovery from https://api.tokenlab.sh/v1/models.
  • Model detail lookup from /v1/models/{model}.
  • Pricing lookup from /v1/models/{model}/pricing.
  • Agent-readable API overview from https://api.tokenlab.sh/llms.txt.
  • Endpoint-family guidance for OpenAI-compatible chat, Responses, Anthropic Messages, Gemini, media, audio, embeddings, rerank, and translation.

​
Install From GitHub

Clone the public repository and install dependencies: 
git clone https://github.com/hedging8563/tokenlab-mcp-server.git
cd tokenlab-mcp-server
npm install
npm test
Start the server over stdio: 
npm start
The optional TOKENLAB_API_BASE environment variable defaults to https://api.tokenlab.sh.

​
Install in Codex

Add the public catalog server to your Codex configuration: 
codex mcp add tokenlab-model-catalog -- npx -y @tokenlabai/mcp-server
Start a new Codex session after adding the server. The public catalog tools do not require TOKENLAB_API_KEY.

​
Claude Desktop Configuration

Add the server to your MCP client configuration. Replace the path with your local clone path: 
{
  "mcpServers": {
    "tokenlab-model-catalog": {
      "command": "node",
      "args": ["/absolute/path/to/tokenlab-mcp-server/src/index.js"]
    }
  }
}
Restart the MCP client after saving the configuration.

​
Cursor, Windsurf, And Other MCP Clients

Use the same command and args in any client that supports stdio MCP servers: 
{
  "command": "node",
  "args": ["/absolute/path/to/tokenlab-mcp-server/src/index.js"],
  "env": {
    "TOKENLAB_API_BASE": "https://api.tokenlab.sh"
  }
}
If your client stores named servers, use tokenlab-model-catalog as the server name.

​
Tools

ToolPurpose
list_modelsList public TokenLab models. Optionally filter with recommended_for such as image, video, embedding, rerank, or translation.
get_modelFetch one model’s public details and supported request format.
get_model_pricingFetch one model’s public pricing details.
get_api_overviewFetch TokenLab’s llms.txt overview for agent-readable endpoint guidance.
  1. Call list_models when the user has not named a model.
  2. Use recommended_for for non-chat tasks, such as image, video, music, 3D, TTS, STT, embeddings, rerank, or translation.
  3. Call get_model before building a non-chat request, retrying a failed request, or switching endpoint families.
  4. Call get_model_pricing when model cost affects the user’s choice.
  5. Use get_api_overview when the agent needs a compact map of TokenLab API families.

​
Native Endpoint Guidance

TokenLab supports OpenAI-compatible routes and native endpoint families. The MCP server helps an agent choose the right route before it writes code:
FamilyCommon route
OpenAI-compatible chat/v1/chat/completions
Responses/v1/responses
Anthropic Messages/v1/messages
Gemini native/v1beta/models/{model}:generateContent
Images/v1/images/generations, /v1/images/edits
Video/v1/videos/generations
Music/v1/music/generations
3D/v1/3d/generations
Audio/v1/audio/speech, /v1/audio/transcriptions, /v1/audio/translations
Embeddings and rerank/v1/embeddings, /v1/rerank
Text translation/v1/translations

​
Use It With TokenLab Skills

The MCP server is useful at runtime, while the TokenLab skills repository teaches coding agents how to generate and repair integration code. Use both when available:
  • MCP server: discover current models, pricing, and endpoint details.
  • tokenlab-api-integration skill: generate runnable API examples and handle structured TokenLab errors.
  • tokenlab-model-picker skill: choose strong models for the user’s task.
  • tokenlab-native-endpoints skill: decide when to use Responses, Anthropic Messages, Gemini, media, audio, embedding, rerank, or translation routes.

​
Troubleshooting

Check that the args path points to your local src/index.js file and that npm install completed in the cloned repository.
Verify that the machine can reach https://api.tokenlab.sh/v1/models. If you override TOKENLAB_API_BASE, make sure it does not include a trailing slash.
Ask the agent to call list_models or get_model before hardcoding a model name. Pairing the MCP server with tokenlab-model-picker gives better results.
No. It is intentionally read-only for public catalog, contract, pricing, and overview data. Use the TokenLab API directly for inference.

​
Resources

GitHub Repository

Source code and local setup instructions

TokenLab Skills

Maintained agent skills for TokenLab integrations

Model Catalog API

Public model discovery endpoint

llms.txt

Agent-readable TokenLab API overview

Glama MCP Server

View the verified TokenLab MCP Server listing

Glama Model Explorer

Explore models, pricing, and native endpoint examples

MCP.so Listing

Discover TokenLab in the AI & Agents directory