English | 简体中文

A Model Context Protocol (MCP) server that provides tools for interacting with Sunra.ai services.

npx @sunra/mcp-server --help

git clone https://github.com/sunra-ai/sunra-clients.git
cd sunra-clients/mcp-server
npm install
npm run build

sunra-mcp-server [options]

Options:
  -t, --transport <type>   Transport type: 'stdio' or 'http' (default: stdio)
  -p, --port <number>      Port for HTTP transport (default: 3000)
  -h, --host <string>      Host for HTTP transport (default: localhost)
  --help                   Show this help message

Examples:
  sunra-mcp-server                           # Start with stdio transport
  sunra-mcp-server --transport http          # Start with HTTP transport on port 3000
  sunra-mcp-server -t http -p 8080           # Start with HTTP transport on port 8080

Add the following to your .cursor/mcp.json file:

{
  "mcpServers": {
    "sunra-mcp-server": {
      "command": "npx",
      "args": ["@sunra/mcp-server"],
      "env": {
        "SUNRA_KEY": "${SUNRA_KEY}"
      }
    }
  }
}

Add the following to your Claude Desktop configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "sunra-mcp-server": {
      "command": "npx",
      "args": ["@sunra/mcp-server"],
      "env": {
        "SUNRA_KEY": "${SUNRA_KEY}"
      }
    }
  }
}

• Base Tools: Submit, status, result, cancel, subscribe operations
• Model Management: List, search, and get schema information for AI models
• File Management: Upload files to Sunra.ai
• Authentication: Secure API key management
• Multiple Transports: Supports both stdio (for Claude Desktop) and HTTP (for Cursor)

• submit - Submit a request to a model endpoint
• status - Check the status of a request
• result - Get the result of a completed request
• cancel - Cancel a pending request
• subscribe - Submit and wait for completion

• list-models - List all available models
• search-models - Search for models by name or description
• model-schema - Get input and output schemas for a specific model endpoint

• upload - Upload files to Sunra.ai storage

• set-sunra-key - Configure your Sunra.ai API key

The model-schema tool now accepts a model slug in the format owner/model/endpoint and returns only the input and output schemas:

# Get schema for a specific model endpoint
model-schema --modelSlug "black-forest-labs/flux-kontext-max/text-to-image"

The tool automatically resolves OpenAPI $ref references to provide fully expanded schemas. For example, if the original OpenAPI schema contains:

{
  "schema": {
    "$ref": "#/components/schemas/TextToVideoInput"
  }
}

The tool will resolve this reference and return the actual schema definition:

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "prompt": {
        "type": "string",
        "description": "Text prompt for video generation"
      },
      "duration": {
        "type": "integer",
        "enum": [5, 10],
        "description": "Duration of the video in seconds"
      }
    },
    "required": ["prompt"]
  }
}

The tool handles:

• ✅ Simple references (#/components/schemas/SchemaName)
• ✅ Nested references within objects and arrays
• ✅ Circular references (marked with $circular: true)
• ✅ Missing references (graceful fallback to original $ref)

Response format:

{
  "success": true,
  "modelSlug": "black-forest-labs/flux-kontext-max/text-to-image",
  "owner": "black-forest-labs",
  "model": "flux-kontext-max",
  "endpoint": "text-to-image",
  "inputSchema": {
    "type": "object",
    "properties": {
      "prompt": {
        "type": "string",
        "description": "Text prompt for image generation"
      }
    },
    "required": ["prompt"]
  },
  "outputSchema": {
    "type": "object",
    "properties": {
      "id": {
        "type": "string",
        "description": "Request ID"
      },
      "status": {
        "type": "string",
        "description": "Request status"
      },
      "output": {
        "type": "object",
        "description": "Generated output"
      }
    }
  }
}

npm test

npm run build

npm start

Set your Sunra.ai API key as an environment variable:

export SUNRA_KEY="your-api-key-here"

Or use the set-sunra-key tool at runtime.

To publish to npm:

npm run build
npm publish

For detailed API documentation, see the Sunra.ai API documentation.