Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/simonw/LLM/llms.txt

Use this file to discover all available pages before exploring further.

LLM’s plugin system makes it straightforward to extend the tool far beyond its built-in capabilities. Built on Pluggy, the same plugin framework used by pytest, it allows third-party packages to hook into LLM and contribute new features — from frontier API models to locally-run LLMs, custom CLI commands, tool integrations, and more. Plugins are distributed as standard Python packages on PyPI and take only a single command to install.

​
What plugins can add

Plugins can contribute any combination of the following to LLM:
  • Language models — both remote API-hosted models (Anthropic, Gemini, Mistral, Groq, …) and locally-run models (Ollama, MLX, llama.cpp, GPT4All, …)
  • Embedding models — vector embedding backends such as sentence-transformers, CLIP, and Jina AI
  • CLI commands — new subcommands bolted on to the llm CLI (e.g. llm cluster, llm cmd, llm python)
  • Tools — callable functions that a model can invoke during a conversation (e.g. math evaluation, SQLite queries, web search)
  • Fragment loaders — custom -f prefix:value loaders that pull content into prompts from GitHub repos, PDFs, Hacker News, websites, and more
  • Template loaders — custom -t prefix:value loaders for fetching prompt templates from external sources such as GitHub or the Fabric collection

​
How plugins are distributed

Plugins are standard Python packages. They declare their hooks via entry points in pyproject.toml (or setup.cfg), which Pluggy discovers automatically at runtime. Because they live on PyPI, installing one is as simple as: 
llm install llm-plugin-name
llm install is a thin wrapper around pip install that targets the correct virtual environment — no manual environment activation needed.

Installing Plugins

Install plugins from PyPI, list what’s installed, uninstall, and control which plugins load with LLM_LOAD_PLUGINS.

Plugin Directory

Browse the full catalogue of available plugins by category: local models, remote APIs, tools, embeddings, and more.

Building a Model Plugin

Step-by-step tutorial on writing your first plugin to support a new language model.

Plugin Hooks Reference

Complete reference for every hook LLM exposes — register_models, register_commands, register_tools, and more.

Build docs developers (and LLMs) love

Get started for freeTalk to us