EverOS
An open-source AI agent memory runtime with local Markdown storage, cloud hosting, and integrations for Claude Code, Codex, OpenClaw, and MCP.
EverOS is a self-hosted memory runtime and Python library for developers building AI agents that need to remember across sessions and applications. It stores conversations, files, and task histories as readable Markdown, with local SQLite and LanceDB indexes for retrieval. The code uses the Apache 2.0 license; a managed cloud service is also available.
Markdown files are the source of truth, so you can read, edit, and version memories with Git. File edits sync to the search indexes. Local deployments store these files and indexes on your hardware, while EverOS Cloud provides managed hosting. Both use the same memory format and retrieval engine, and memories can move between them through Markdown export.
The memory layer separates user profiles and past interactions from agent task records and reusable skills. Background consolidation combines related episodes and refines profiles and skills between sessions. Successful task histories can become procedures that other agents reuse. Memory scopes let developers separate shared knowledge from project and session context.
EverOS works with Claude Code, Codex, OpenClaw, Hermes, MCP, and the OpenAI and Anthropic SDKs. It supports keyword search and hybrid keyword/vector retrieval; embeddings and reranking require providers. Multimodal ingestion covers images, PDFs, audio, office files, and URLs. Its Knowledge Wiki adds editable Markdown pages linked to source material, with topic search and APIs for managing entries.
