Repolode
EverMind-AI

EverMind-AI/EverOS

A local-first, Markdown-native long-term memory layer for AI agents, storing memories as editable files with SQLite and LanceDB indexes.

Stars
13,525
Forks
1,074
Open issues
58
Last commit
18 hours ago

What is it?

EverOS is a Python library and memory runtime that gives AI agents persistent memory across coding assistants, apps, and workflows. Conversations, files, and agent trajectories are stored as readable Markdown files, which act as the source of truth and can be edited directly, diffed, and versioned with Git.

A local SQLite and LanceDB index layer keeps the Markdown files searchable. Memories are separated into user tracks (episodes and profiles) and agent tracks (cases and skills), and retrieval can be scoped by user, agent, app, project, and session IDs. An offline reflection step merges episode clusters and refines profiles and skills between sessions.

The server exposes a versioned HTTP API under /api/v2. A minimal setup needs only an OpenRouter API key for add, flush, and keyword search. Embedding, rerank, and multimodal providers are optional and unlock hybrid search, Knowledge Wiki, and image, PDF, audio, or office-file ingestion.

Key features

  • 01Stores memories as editable, Git-versionable Markdown files.
  • 02Indexes content locally with SQLite and LanceDB, with no external database required.
  • 03Keeps separate user episodes/profiles and agent cases/skills.
  • 04Scopes retrieval by user, agent, app, project, and session IDs.
  • 05Runs offline reflection to merge episode clusters and refine profiles and skills.
  • 06Exposes a versioned HTTP API under /api/v2 and includes a standalone demo and CLI.

Use cases

  • Giving a coding assistant such as Claude Code or Cursor memory of past project sessions via MCP.
  • Adding long-term user memory to a chatbot or voice assistant backend.
  • Keeping auditable, human-editable agent memory in plain Markdown files under Git.
  • Integrating persistent memory into workflow tools like Dify through its plugin.

Quick start

  1. 1

    Install EverOS with uv

    uv pip install everos
  2. 2

    Try the standalone demo (no key required)

    everos demo
  3. 3

    Initialize config and add your OpenRouter API key to ~/.everos/everos.toml

    everos init
  4. 4

    Start the EverOS server

    everos server start

Things to know

  • Full memory search needs an OpenRouter API key; hybrid and vector features need an extra embedding provider.
  • Office document ingestion requires LibreOffice installed on the system.
  • Setup involves several optional providers and config files, so self-hosting takes some effort.

Topics

FAQ

What is EverOS?
A local-first, Markdown-native long-term memory layer for AI agents, storing memories as editable files with SQLite and LanceDB indexes. EverOS is a Python library and memory runtime that gives AI agents persistent memory across coding assistants, apps, and workflows. Conversations, files, and agent trajectories are stored as readable Markdown files, which act as the source of truth and can be edited directly, diffed, and versioned with Git.
Is EverOS free to use? What license does it have?
EverOS is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is EverOS still maintained?
Actively maintained. The last commit to EverOS was on Oct 20, 2026.
What can I use EverOS for?
Giving a coding assistant such as Claude Code or Cursor memory of past project sessions via MCP. Adding long-term user memory to a chatbot or voice assistant backend. Keeping auditable, human-editable agent memory in plain Markdown files under Git. Integrating persistent memory into workflow tools like Dify through its plugin.
How do I install EverOS?
Install EverOS with uv: uv pip install everos Try the standalone demo (no key required): everos demo Initialize config and add your OpenRouter API key to ~/.everos/everos.toml: everos init Start the EverOS server: everos server start
Where can I get EverOS?
The source code and installation instructions are on GitHub: https://github.com/EverMind-AI/EverOS

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

Related projects