Local LLM
marm-memory
A local-first 3-in-1 AI memory layer and MCP server.
★ 416 starsApache-2.0
View guide →A local-first context and memory layer for your AI agents.
zhigeng is an open-source local llm project by Littlesheepxy (Littlesheepxy/zhigeng on GitHub). It has 490 GitHub stars and 20 forks, is released under the NOASSERTION license, and was last updated on GitHub on 1 September 2026.
Read the setup at zhigeng.app.
Get the macOS app.
Bring your own model keys.
Voice in, contextual memory out.
Agents remember across sessions.
Talk to your AI on Mac.
Routes to Codex or Claude Code.
A pocket memory terminal.
Your context stays yours.
Use your own model keys.
Download the source and enter the folder.
git clone https://github.com/Littlesheepxy/zhigeng.git
cd zhigengInstall the project dependencies.
npm installLaunch per the README guide.
npm run devA local-first context and memory layer for AI on Mac and iOS, at Littlesheepxy/zhigeng.
Bring your own key: you supply the model API keys.
Yes. It is local-first by design.
The code is on GitHub; the license is source-available noncommercial, so check terms.
Codex and Claude Code, among others.
A local-first 3-in-1 AI memory layer and MCP server.
A memory operating system for long-term, personalized AI agents.
SillyTavern extension giving AI roleplay characters long-term memory.
Local-first personal knowledge layer for AI agents.
A local-first OS for building and running AI agents.
Local-first inbox designed for long-running AI agents, built.
The most-starred other Local LLM projects in the LibreAI catalogue — the realistic shortlist if zhigeng is not the right fit.
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.
Free and open — grab the source, star the repo, and start creating.