MenteDB is a Rust database for AI agents. It is described as memory for AI, built from scratch to pre-digest knowledge for single-pass transformer consumption rather than acting as a wrapper around another system. The page frames it as a response to an AI memory problem: current systems take one context window, cannot re-read or reflect, and may not recognize what they are missing.
Its core functions center on extracting and managing memory across conversations and sessions. MenteDB says it automatically extracts facts, preferences, decisions, and corrections, then stores them for later retrieval. It also offers semantic search across sessions, contradiction detection, pain warnings that surface before a mistake is repeated, and cloud sync so every device sees the same memories. The feature list further includes background enrichment that extracts semantic facts, links entities, detects communities, and builds user profiles; continuous stream processing with real-time belief updates; write-time inference; trajectory tracking; phantom memories for knowledge gaps; interference shielding; pain signals; speculative pre-assembly; bi-temporal validity with point-in-time queries; LLM cognitive inference; and hybrid search combining BM25, HNSW vector search, and RRF fusion.
The tool is presented for AI agents and developers building memory into agent workflows. The page says Claude Code hooks make memory automatic on every turn, and that ChatGPT, Cursor, Copilot, and the claude.ai connector can connect with one URL. It also lists native SDKs for Rust, Python, and TypeScript, with installation examples for pip, npm, and cargo. The page shows code for use in an agent turn and says one call per conversation turn handles extraction, storage, retrieval, contradiction detection, and background enrichment automatically.
MenteDB is offered as open source under the Apache 2.0 license and is marked beta. The site also includes an API key management dashboard with scoped access control, webhooks, and delivery logs.
MenteDB is a Databases (SQL, NoSQL, vector, graph) project. It focuses on providing AI agents with persistent, queryable long-term memory instead of relying on static context windows. It is built as an open-source project for AI agent developers. The project is open source (Apache-2.0). It runs on the web, the command line, and API, and it can be self-hosted.
Behind MenteDB is MenteDB, and it first shipped in 2026. Development happens publicly on GitHub with 98 stars and 246 commits in the last 90 days. Key capabilities include Long-term Memory, Single-pass Consumption, and Semantic Search. It exposes integrations via a public API. MenteDB is currently in beta.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do