🦙 ollama-mcpo-adapter

🦙 ollama-mcpo-adapter

Expose MCPO, the MCP-to-OpenAPI proxy server, tools as Ollama-compatible functions using a simple Python adapter and optional runtime service.

---

✨ Features

• 🔌 Connect to a MCPO instance
• ⚙️ Launch your own MCPO server programmatically via MCPOService
• 🔁 List MCP tools exposed via OpenAPI as Ollama-compatible tool functions

---

🚀 Quickstart

1. Install

pip install ollama-mcpo-adapter

Usage with Existing MCPO Instance

Assuming you have MCPO running like this:

uvx mcpo --port 5090 --config /path/to/mcp_config.json

You can get all available functions in Ollama ToolCall format with the adapter:

from ollama_mcpo_adapter import OllamaMCPOAdapter

adapter = OllamaMCPOAdapter(host="localhost", port=5090, config_path="/path/to/mcp_config.json")
# Gets tool descriptions from MCPO FastAPI /docs
tools = adapter.list_tools_ollama()

You can omit the config path. But discovery of MCP server names is more reliable with a provided config. Otherwise, the server names will be read from the automatically generated OpenAPI docs MCPO provides which might change in the future.

---

Usage with Local MCPO Service

You can start a MCPO service with this extension:

from ollama_mcpo_adapter import MCPOService

# Provide your mcp config as JSON file or dictionary
mcp_config = {
    "mcpServers": {
        "time": {"command": "uvx", "args": ["mcp-server-time", "--local-timezone=Europe/Berlin"]}
    }
}
mcpo = MCPOService("127.0.0.1", 4090, config=mcp_config,
                   # -OR- from an existing mcp_config file 
                   config_path="path/to/mcp_config.json")
# MCPOSService class handles MCPO server start-up and shutdown and in a subprocess
mcpo.start(wait=True)
...
mcpo.stop()

Then get all available tools with the adapter:

from ollama_mcpo_adapter import OllamaMCPOAdapter
adapter = OllamaMCPOAdapter("127.0.0.1", 4090)
tools = adapter.list_tools_ollama()

Send this to Ollama:

from ollama import Client
# Send a prompt to Ollama using discovered tools
client = Client(host="http://127.0.0.1:11434")
response = client.chat(
    model="qwen2.5-coder:14b-instruct-q4_K_M",
    messages=[{"role": "user", "content": "Write a file..."}],
    tools=tools,
)

And finally call the tools:

# Handle any tool calls
if response.message.tool_calls:
    adapter.call_tools_from_response(response.message.tool_calls)

---

Env

• MS Windows npx path, you can overwrite npx with a path to npx in the config parser

  example: WIN_NODEJS_NPX_PATH=C:\Program Files\nodejs\npx.cmd

---

🧪 Running Tests

pytest

---

📂 Project Structure

ollama_mcpo_adapter/
├── adapter.py        # Tool discovery + Ollama integration
├── service.py        # Optional: launch MCPO programmatically
├── service_runner.py # MCPO subprocess control
├── config_parser.py  # MCP config parsing helpers
├── dispatcher.py     # Dispatch tool calls

---

📜 License

MIT. See LICENSE.