Rememori
Provides an embedded memory engine with remember, recall, and forget operations, using embeddings, tags, entities, importance, and time decay for semantic and graph-based retrieval.
rememori is an embedded memory engine for AI agents, written in pure TypeScript with zero dependencies, that solves the problem of agents forgetting context across sessions by providing a simple API to store, recall, and forget memories with semantic search and entity-based graph recall. It targets developers building AI agents, including support bots, coding agents, and local-first apps, and is positioned as a lightweight alternative to vector databases and embedding pipelines, running anywhere JavaScript runs, from Node.js to browsers and edge workers. It is delivered as an npm package and an MCP server, with optional local embedding via Ollama.
Key features
- Pure TypeScript, zero dependencies
- Embedded memory engine
- Remember, recall, forget API
- Semantic similarity ranking
- Entity-based graph recall
- Importance and time decay
- Reinforce and demote memory
- Pluggable embedders (Ollama, OpenAI-compatible)
- Runs in Node, Bun, Browser, Edge
- MCP server support
- No social media activity detected
- No GTM activity detected
- Software developers
- Engineering teams
- Startups