Ifly Spark Agent MCP

What is Ifly Spark Agent MCP about?

A lightweight MCP server that connects to the iFlytek SparkAgent platform, allowing clients to trigger the platform’s task chain. The server automatically registers a tool called upload_file, which accepts a file path and forwards the file to the SparkAgent service.

How to use Ifly Spark Agent MCP?

Local debugging

# stdio transport (default)
uv run ifly-spark-agent-mcp

# SSE transport on a custom port
uv run ifly-spark-agent-mcp --transport sse --port 8000

With an MCP client (Python example)

import asyncio
from mcp.client.session import ClientSession
from mcp.client.stdio import StdioServerParameters, stdio_client

async def main():
    async with stdio_client(
        StdioServerParameters(command="uv", args=["run", "ifly-spark-agent-mcp"])
    ) as (read, write):
        async with ClientSession(read, write) as session:
            await session.initialize()
            tools = await session.list_tools()
            print(tools)
            result = await session.call_tool("upload_file", {"file": "/path/to/file"})
            print(result)

asyncio.run(main())

Configuration for persistent MCP clients (e.g., Claude Desktop)

Add an entry to claude_desktop_config.json or mcp.json:

{
  "mcpServers": {
    "ifly-spark-agent-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "/path/to/ifly-spark-agent-mcp",
        "run",
        "ifly-spark-agent-mcp"
      ],
      "env": {
        "IFLY_SPARK_AGENT_BASE_URL": "xxxx",
        "IFLY_SPARK_AGENT_APP_ID": "xxxx",
        "IFLY_SPARK_AGENT_APP_SECRET": "xxxx"
      }
    }
  }
}

Or using uvx directly from the GitHub repository:

{
  "mcpServers": {
    "ifly-spark-agent-mcp": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/iflytek/ifly-spark-agent-mcp",
        "ifly-spark-agent-mcp"
      ],
      "env": {
        "IFLY_SPARK_AGENT_BASE_URL": "xxxx",
        "IFLY_SPARK_AGENT_APP_ID": "xxxx",
        "IFLY_SPARK_AGENT_APP_SECRET": "xxxx"
      }
    }
  }
}

Key features of Ifly Spark Agent MCP

• Simple command‑line start with stdio or Server‑Sent Events transport.
• Auto‑registration of an upload_file tool for seamless file submission.
• Works with any MCP‑compatible client (Python SDK, Claude Desktop, etc.).
• Environment‑driven configuration for SparkAgent credentials and endpoint.

Use cases of Ifly Spark Agent MCP

• Embedding file‑based workflows into AI agents that need to process documents via SparkAgent.
• Rapid prototyping of custom agents that call SparkAgent services without writing HTTP code.
• Integrating SparkAgent capabilities into Claude or other LLM‑driven assistants.

FAQ from the Ifly Spark Agent MCP

Q: Which transport should I use? A: The default stdio transport works for most local development. Use SSE (--transport sse) when you need a persistent HTTP connection.

Q: How do I set the SparkAgent credentials? A: Provide IFLY_SPARK_AGENT_BASE_URL, IFLY_SPARK_AGENT_APP_ID, and IFLY_SPARK_AGENT_APP_SECRET as environment variables in the server configuration.

Q: Can I call other SparkAgent APIs? A: Currently the server only exposes the upload_file tool, but you can extend it by adding more tools in the source code.

Q: Is there a Docker image? A: Not provided in the repository; you can containerize the project by installing the Python dependencies and running the uv run command.

ifly-spark-agent-mcp

This is a simple example of using MCP Server to invoke the task chain of the iFlytek SparkAgent Platform.

Usage

Local debugging

Start the server using either stdio (default) or SSE transport:

# Using stdio transport (default)
uv run ifly-spark-agent-mcp

# Using SSE transport on custom port
uv run ifly-spark-agent-mcp --transport sse --port 8000

By default, the server exposes a tool named "upload_file" that accepts one required argument:

• file: The path of the uploaded file

MCP Client Example

Using the MCP client, you can use the tool like this using the STDIO transport:

import asyncio
from mcp.client.session import ClientSession
from mcp.client.stdio import StdioServerParameters, stdio_client


async def main():
    async with stdio_client(
        StdioServerParameters(command="uv", args=["run", "ifly-spark-agent-mcp"])
    ) as (read, write):
        async with ClientSession(read, write) as session:
            await session.initialize()

            # List available tools
            tools = await session.list_tools()
            print(tools)

            # Call the upload_file tool
            result = await session.call_tool("upload_file", {"file": "/path/to/file"})
            print(result)


asyncio.run(main())

Usage with MCP client

Use on Claude

To add a persistent client, add the following to your claude_desktop_config.json or mcp.json file:

1. Use uv

{
  "mcpServers": {
    "ifly-spark-agent-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "/path/to/ifly-spark-agent-mcp",
        "run",
        "ifly-spark-agent-mcp"
      ],
      "env": {
        "IFLY_SPARK_AGENT_BASE_URL": "xxxx",
        "IFLY_SPARK_AGENT_APP_ID": "xxxx",
        "IFLY_SPARK_AGENT_APP_SECRET": "xxxx"
      }
    }
  }
}

2. Use uvx with github repository

{
    "mcpServers": {
        "ifly-spark-agent-mcp": {
            "command": "uvx",
            "args": [
                "--from",
                "git+https://github.com/iflytek/ifly-spark-agent-mcp",
                "ifly-spark-agent-mcp"
            ],
            "env": {
              "IFLY_SPARK_AGENT_BASE_URL": "xxxx",
              "IFLY_SPARK_AGENT_APP_ID": "xxxx",
              "IFLY_SPARK_AGENT_APP_SECRET": "xxxx"
            }
        }
    }
}