Sara icon
Sara icon

Sara

 Like

Sara is an MIT-licensed AI companion and coding assistant that runs on your own server. It combines a customizable personality and editable persistent memory with planning, document reading and coding project workspaces.

Coding workspace in Sara using a local model on a throwaway test server.

Cost / License

  • Free
  • Open Source (MIT)

Application type

Since

  • New

Updated

  • Updated Oct 2026

Platforms

  • Self-Hosted  Requires a Sara server and a configured local or cloud model. Native server requires Node.js 24+; Docker Compose is also documented.
  • Windows  Windows 10/11 (64-bit); installer available from GitHub releases.
  • Android
0likes
0articles
Save

Features

Properties

Features

Highlights All activities

Sara News & Activities

Recent activities

  • Backstage-Coding added Sara
  • Replika icon
    Backstage-Coding added Sara as alternative to Replika

Sara information

AlternativeTo Categories

Development, AI Tools & Services, Social & Communications

GitHub repository

  •  2 Stars
  •  2 Forks
  •  6 Open Issues
  •   Updated  
View on GitHub

Popular alternatives

View all
Sara was added to AlternativeTo by Backstage-Coding on and this page was last updated .

No comments or reviews, maybe you want to be first?

Official Links

What is Sara?

Sara is an MIT-licensed AI companion and coding assistant that runs on your own server. It combines a customizable personality and editable persistent memory with planning, document reading and coding project workspaces.

For development work, Sara can edit files, use Git and run commands and tests, with approval controls for risky actions. It supports local models through Ollama, vLLM and llama.cpp, as well as supported cloud providers. A web interface and desktop and mobile clients connect to your server.

The core is written in TypeScript and runs on Node.js 24 or newer with no runtime dependencies. Install using the documented native or Docker Compose setup. Users provide their own hardware and models; cloud model services may charge separately. Capabilities and results depend on the selected model and deployment environment.