Repolode
Submit
sui-ni2

sui-ni2/personal-ai-os

A local-first, provider-neutral personal AI workspace with chat, reviewed memory, auditable tools, MCP support, and isolated project plugins.

Open-source alternative to:ChatGPT

View on GitHub
Stars
58
Forks
15
Open issues
7
Last commit
21 hours ago

What is it?

Personal AI OS is a self-hosted workspace for long-running AI work. Projects, conversation context, reviewed memory, tools, and execution records are stored in the workspace rather than inside any one model provider, so users can switch between OpenAI, Anthropic, or a local Ollama model without moving their project data.

The app is built with a Next.js front end and a FastAPI backend using SQLite. It streams responses over SSE, keeps conversation history across restarts, and records tool calls as observable traces. Tools are reached through an allowlisted MCP gateway, and the model cannot submit shell commands. Secrets remain in server-side environment variables and are never returned by the settings view.

The project starts without any provider key, can be run with Docker Compose bound to localhost, and includes a Windows bootstrap script. It also ships with a General project, a Soccer plugin, and an isolated P5 / 排列5 plugin that uses the same project contract.

Key features

  • 01Chat with OpenAI, Anthropic, or optionally enabled local Ollama models behind one streaming interface.
  • 02Restart-safe conversation history with project filtering and visible execution traces.
  • 03Allowlisted MCP gateway with HTTP and fixed-command stdio connectors.
  • 04Installable mobile PWA shell with optional single-user password gate.
  • 05Redacted diagnostics and secrets kept server-side.

Use cases

  • Running a self-hosted AI assistant that keeps project context independent of any single model vendor.
  • Experimenting with several LLM providers while keeping conversation and memory data local.
  • Connecting AI chat to your own tools through a controlled MCP allowlist.
  • Trying the app without paying for an API key using the no-key Docker startup path.

Topics

#personal-ai#llm#self-hosted#mcp#ollama#openai#anthropic#privacy#ai-workbench#developer-tools

FAQ

What is personal-ai-os?
A local-first, provider-neutral personal AI workspace with chat, reviewed memory, auditable tools, MCP support, and isolated project plugins. Personal AI OS is a self-hosted workspace for long-running AI work. Projects, conversation context, reviewed memory, tools, and execution records are stored in the workspace rather than inside any one model provider, so users can switch between OpenAI, Anthropic, or a local Ollama model without moving their project data.
Is personal-ai-os free to use? What license does it have?
personal-ai-os is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is personal-ai-os still maintained?
Actively maintained. The last commit to personal-ai-os was on Oct 21, 2026.
What can I use personal-ai-os for?
Running a self-hosted AI assistant that keeps project context independent of any single model vendor. Experimenting with several LLM providers while keeping conversation and memory data local. Connecting AI chat to your own tools through a controlled MCP allowlist. Trying the app without paying for an API key using the no-key Docker startup path.
Where can I get personal-ai-os?
The source code and installation instructions are on GitHub: https://github.com/sui-ni2/personal-ai-os

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects