LLM locaux
Web LLM
High-performance In-browser LLM Inference Engine
★ 19,2 k étoilesApache-2.0
Voir le guide →WebMCP est un projet de LLM locaux open source, créé par jasonjmcghee (jasonjmcghee/WebMCP sur GitHub). Il compte 883 étoiles sur GitHub et 54 forks, il est publié sous la licence MIT et a été mis à jour pour la dernière fois sur GitHub le 15 février 2026.
Early proposal and implementation for connecting LLMs to web pages.
Understand the original WebMCP idea.
See how the early implementation worked.
Look at how the concept evolved.
Use it as inspiration, not infrastructure.
One of the first WebMCP attempts.
A working sketch of the idea.
About LLMs interacting with pages.
Transparent evolution of the concept.
Free to study and reuse.
Useful for standards archaeology.
Download the source and move into the project folder.
git clone https://github.com/jasonjmcghee/WebMCP.git
cd WebMCPEvery project documents its own setup; start with install requirements.
cat README.md | head -60Follow the project's launch command — Docker, pip, npm, or a binary release.
# see README for the exact launch commandNo, the author points to successor work by others.
Better to follow the current evolution of the idea instead.
MIT licensed — commercial use allowed.
Work continuing under the broader web machine learning community.
It is an honest look at where the web-AI connection idea started.
High-performance In-browser LLM Inference Engine
Typescript/React Library for AI Chat 💬🚀
⚡ Pure-Rust WebGPU inference engine — OpenAI-API compatible, GGUF native, runs on any GPU. No Python. No llama.cpp. Single binary.
Fast Rust scraper that turns web pages into clean data for LLMs.
Run local LLMs like Llama and DeepSeek right inside your web browser.
A lightweight vLLM implementation built from scratch in about 1,200 readable lines of Python.
Les autres projets de la catégorie LLM locaux avec le plus d'étoiles du catalogue LibreAI — la sélection réaliste d'alternatives si WebMCP ne vous convient pas.
The simplest way to run open language models on your own machine.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
LLM inference in C/C++
A high-throughput and memory-efficient inference and serving engine for LLMs
🤯 LobeHub is your Chief Agent Operator, organizing your agents into 7×24 operations by hiring, scheduling, and reporting on your entire AI team.
Projets de la catégorie LLM locaux proches dans le catalogue, les plus proches en étoiles de WebMCP — continuez à explorer de part et d'autre.
Browse and resume your AI coding-agent sessions on macOS.
Local security auditor for AI API relays — catches injection and swaps.
A talking LLM voice assistant that runs fully offline on your computer.
Desktop AI agent platform with one-click models and offline-first RAG.
One-command installer for a full self-hosted AI stack on your server.
Local-first desktop workspace designed for running.
Gratuit et ouvert : récupérez le code source, mettez une étoile au dépôt et commencez à créer.