Local-first persistent memory for AI coding agents, storing decisions and lessons in SQLite and serving them over MCP to Claude Code, Cursor and other clients.
ELY-Open-Source/Elysian
A local memory store for AI that records who each memory belongs to, its source, when it was true, and who may read it.
Open-source alternative to:Mem0
- Stars
- 13
- Forks
- 0
- Open issues
- 0
- Last commit
- 7 days ago
What is it?
Elysian is a local memory store for AI agents and assistants. Each memory is stored with its holder, origin, scenario, event time, validity window, visibility and retention settings, so a store shared by several people or agents keeps their answers separate. Retrieval answers as the person asking, and when the store lacks an answer it reports who is likely to know rather than guessing.
It ingests text, PDFs, code, conversations, pictures, screenshots and video (with speech transcription), and searches them with a words index plus optional dense vectors, picture search and a local model reranker. Writes go through one set of rules exposed over HTTP, a CLI and MCP, so coding agents can read and write the same store. Context Cartridges are scoped memory packages that can be mounted to limit searches to one project.
The repository includes a benchmark suite with reports for recall, latency, speech and picture retrieval, and comparisons against Mem0. Some generation and training features are written but not yet run in the repository, and the README marks them as not measured.
Key features
- 01Stores each memory with holder, origin, validity period and visibility settings.
- 02Answers searches from the asker's point of view and names who knows when it has no answer.
- 03Runs local, with SQLite holding atoms, full-text index and vectors in one file.
- 04Exposes remember, supersede, forget and search over HTTP, CLI and MCP.
- 05Provides an MCP server so coding agents such as Claude Code and Codex can share the store.
- 06Ingests PDFs, code, conversations, pictures and video with speech transcription.
Use cases
- Giving several family members or team members separate, private AI memory on one machine.
- Letting coding agents across different tools read and write one project memory store.
- Searching personal documents, screenshots and recordings locally without sending them to a cloud service.
- Scoping an AI assistant's answers to a single project by mounting its memory package.
Quick start
- 1
Install the memory store, search, API and CLI from the backend folder
pip install ./backend - 2
Store a memory for a holder
elysian remember "the archive key is in the second drawer" --holder wife --scenario home - 3
Search the store as a holder
elysian search "where did we put the archive key" --as-holder wife
Things to know
- The generator features, such as trained LoRA adapters and visual multi-hop addressing, have not been run and need a CUDA machine and a model download.
- The full local model stack (embeddings, CLIP, speech, training) needs the optional models extra and is heavy to install.
In collections
Topics
FAQ
- What is Elysian?
- A local memory store for AI that records who each memory belongs to, its source, when it was true, and who may read it. Elysian is a local memory store for AI agents and assistants. Each memory is stored with its holder, origin, scenario, event time, validity window, visibility and retention settings, so a store shared by several people or agents keeps their answers separate. Retrieval answers as the person asking, and when the store lacks an answer it reports who is likely to know rather than guessing.
- Is Elysian free to use? What license does it have?
- Elysian is open source under the MIT license. Check the license terms before using it in commercial products.
- Is Elysian still maintained?
- Actively maintained. The last commit to Elysian was on Oct 31, 2026.
- What can I use Elysian for?
- Giving several family members or team members separate, private AI memory on one machine. Letting coding agents across different tools read and write one project memory store. Searching personal documents, screenshots and recordings locally without sending them to a cloud service. Scoping an AI assistant's answers to a single project by mounting its memory package.
- How do I install Elysian?
- Install the memory store, search, API and CLI from the backend folder: pip install ./backend Store a memory for a holder: elysian remember "the archive key is in the second drawer" --holder wife --scenario home Search the store as a holder: elysian search "where did we put the archive key" --as-holder wife
- Where can I get Elysian?
- The source code and installation instructions are on GitHub: https://github.com/ELY-Open-Source/Elysian
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Local-first persistent memory for AI agents, stored as markdown files with a knowledge graph, retrieval, and an MCP server.
A local Rust tool that gives Claude Code persistent memory of past conversations via semantic search, exposed as an MCP server.
Local-first AI memory system that stores conversation history verbatim and retrieves it with semantic search, exposed as an MCP server and CLI.
Long-term memory for coding agents that stores memories in a local SQLite database and shares them across Claude Code, Codex, Cursor and other MCP clients.
Local-first persistent memory for AI coding agents like Claude Code, Cursor and Codex, served over MCP from a SQLite file on your disk.