OpenRouter llm models API

Models served by OpenRouter: context window, modality, dollars per million tokens and accepted parameters.

Ask your agent

“Call openrouter/llm-models with no arguments and summarize what comes back.”

Works in any MCP client connected to https://data.openkrill.app/mcp. Install in one line.

Path: openrouter/llm-models
MCP tool: call_endpoint
Credits: 1
Cache: 30 min

Parameters

querystring, optional	Free text matched against model ids and names, such as claude or gpt
providerstring, optional	Only models from this provider, such as openai or anthropic
min_contextinteger, optional	Only models with at least this many context tokens
limitinteger, optional	1-100 models, default 20

Response

data.resources[] uses the same keys as every other platform: author, engagement, metadata, with OpenRouter-only fields under extras. A failed upstream call is never billed.

MCP tool call

{
  "name": "call_endpoint",
  "arguments": {
    "path": "openrouter/llm-models",
    "params": {}
  }
}

curl

curl "https://data.openkrill.app/v1/openrouter/llm-models" \
  -H "x-api-key: YOUR_API_KEY"

Response

{
  "success": true,
  "platform": "openrouter",
  "endpoint": "openrouter/llm-models",
  "data": {
    "resources": [
      { "id": …, "title": …, "text": …, "url": …,
        "author": {…}, "engagement": {…},
        "metadata": {…}, "extras": {…} }
    ]
  },
  "credits_used": 1,
  "cached": false
}