Локальные LLM
exo
Run frontier AI locally.
★ 47,7 тыс. звёздApache-2.0
Смотреть руководство →agent-safehouse — проект с открытым исходным кодом из категории «Локальные LLM» от eugene1g (eugene1g/agent-safehouse на GitHub). У него 2 092 звезды на GitHub и 99 форков, он распространяется по лицензии Apache-2.0 и последний раз обновлялся на GitHub 30 сентября 2026.
Sandbox for local AI agents so they only touch what they need.
Set it up on your machine.
Choose which paths the agent may access.
Run your coding agent inside the sandbox.
The agent operates within its allowed scope.
Agents see only approved paths.
Scope exactly what agents may touch.
Built for agents running on your machine.
Limits damage from agent errors.
Reduces prompt-injection blast radius.
Works with Claude Code on Mac.
Download the source and move into the project folder.
git clone https://github.com/eugene1g/agent-safehouse.git
cd agent-safehouseEvery 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 commandSimilar idea: isolation, but scoped to what an AI agent may read and write.
Apache 2.0 licensed — commercial use allowed with attribution.
Claude Code on macOS is the headline; check the docs for others.
Sandboxing adds some overhead; check the docs for details.
No sandbox is perfect; treat it as risk reduction, not a guarantee.
Run frontier AI locally.
Hundreds of models & providers. One command to find what runs on your hardware.
Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU.
Local-first dashboard of tokens, models, and cost for coding agents.
AI OSINT agent with 20 tools and REPL — for authorized research only.
A local-first context and memory layer for your AI agents.
Другие проекты из категории «Локальные LLM» с наибольшим числом звёзд в каталоге LibreAI — реалистичный список альтернатив, если agent-safehouse вам не подходит.
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.
Соседние проекты из категории «Локальные LLM» в каталоге, ближайшие по числу звёзд к agent-safehouse — продолжайте изучать полку по обе стороны от него.
Local-first desktop where a commander LLM directs specialist coding agents.
Local API bridge to run Claude and Antigravity inside other coding clients.
Run AI models locally on your machine with node.js bindings for llama.cpp. Enforce a JSON schema on the model output on the generation level
StepFun's open 196B MoE model that activates 11B parameters per token, built for fast agentic reasoning.
Let any LLM actually watch a video — frames plus transcript, run locally.
One shared local memory hub so every AI agent remembers the same you.
Бесплатно и открыто — забирайте исходный код, ставьте звезду репозиторию и начинайте создавать.