Skip to main content

llm-tools-mcp

mcp-logo

Connect to MCP servers right from your shell. Plugin for llm (by @simonw).

PyPI Changelog Tests License

demo-long

[!Note] Current focus: Authorization #4

Installation

Install this plugin in the same environment as LLM:

llm install llm-tools-mcp

Usage

[!WARNING] It's recommended to use the --ta flag and approve each tool execution.

  1. Create mcp.json file in ~/.llm-tools-mcp.

    Example file:

    {
      "mcpServers": {
        "filesystem": {
          "command": "npx",
          "args": [
            "-y",
            "@modelcontextprotocol/server-filesystem",
            "~/demo"
          ]
        }
      }
    }
    
  2. List available tools.

    [!WARNING] llm tools is not implemented for dynamic toolboxes (see this), so tools loaded as part of MCP won't be visible.

    llm tools list
    
  3. Run llm with tools.

    llm --ta -T MCP "what files are in the demo directory? show me contents of one of the files (any)"
    

Other examples

Dynamically change your MCP config:

llm --ta -T 'MCP("/path/to/custom/mcp.json")' "your prompt here"

Development

Now (to be verified)

  • Sync dependencies: uv sync --all-extras
  • Run linters / type checker: ./check.sh
  • Run tests: ./test.sh
  • Run end to end tests: ./e2e/e2e-docker.sh

To Do

  • Release alpha version
  • support all transports
    • streamable http
    • sse
    • stdio
  • Build a solid test suite
    • test config file validation
    • test sse with dummy server
    • test stdio with dummy server
    • test http streamable with dummy server (see #1)
    • manual test for sse with real server
    • manual test for stdio with real server
    • manual test for http streamable with real server
  • Redirect stdout/stderr from the MCP SDK to a file or designated location
  • Reuse stdio connections
  • Support non-stdio MCP servers
  • Handle tool name conflicts (prefix with mcp server name?)
  • Gather feedback on the ~/.llm-tools-mcp directory naming
  • Improve failure handling:
    • When connecting to an MCP server fails
    • When mcp.json is malformed
  • Improve this README:
    • Add more detail in the Development section (mention uv?)

Metadata

Release files for llm-tools-mcp 0.4

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for llm-tools-mcp 0.4
File Size Uploaded
llm_tools_mcp-0.4.tar.gz 13.2 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for llm-tools-mcp 0.4
File Interpreter ABI Platform
llm_tools_mcp-0.4-py3-none-any.whl Python 3 none any Details

Total release size: 24.2 kB

Release files / llm_tools_mcp-0.4.tar.gz

Download URL llm_tools_mcp-0.4.tar.gz
Size 13.2 kB
Tags Source
SHA-256 checksum
How to use checksums
a893144021c026ade06c878e3b273bcf18f67e46d172e71bf86e2ed42093561e
BLAKE2b-256 checksum
How to use checksums
5d4aabc038b896b9ee67b65305531820f2a3b9c499a21d76f5c277c45b7dfb32
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.7

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 29, 2025.

Transparency log

Release files / llm_tools_mcp-0.4-py3-none-any.whl

Download URL llm_tools_mcp-0.4-py3-none-any.whl
Size 11.0 kB
Tags Python 3
SHA-256 checksum
How to use checksums
a3907ebb14fb9e7a431c360602e9bcf767b43f5bbff2f8f7cc395544195a32a2
BLAKE2b-256 checksum
How to use checksums
e3c3e136286bd0589ccd9b2cff56d7f8778dcf78aa3894b82a726434bbeac1db
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.7

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 29, 2025.

Transparency log
Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page