Mnemolith

Mnemolith is a shared local memory system for AI coding tools that keeps information synchronized with code changes.

Mnemolith provides a shared local memory layer for AI coding tools like Claude Code, Cursor, and Codex. It links memories directly to project files using git content hashes, automatically retiring stale information when the underlying code changes. The software runs entirely on your local machine using a custom database engine with an HNSW vector index and MCP server support. It handles conflict review, duplicate merging, and full memory provenance. Users can also import and export notes from files like CLAUDE.md and AGENTS.md to maintain consistent context across sessions.

Mnemolith at a glance

Best for: Software engineer
Runs on: macOS, Linux, Windows, API, self-hosted, open source
Pricing: Free
Released: October 2026
Category: AI coding tool
Website: github.com/anishchugh05/mnemolith

Key features

Retires stale memories automatically when linked files change
Shares context across Claude Code, Cursor, and Codex via MCP server
Performs semantic search using an HNSW vector index
Merges near-identical duplicates and flags contradictions
Syncs current memories into AGENTS.md and CLAUDE.md files
Provides a local dashboard UI to review and manage memories
Maintains full provenance and history for every saved fact

Use it for

Remembering project-specific test configurations and setup steps
Syncing AI memory across multiple coding assistants in a repository
Importing existing notes from markdown files into structured memory
Detecting and resolving contradictory facts in project documentation

Alternatives to Mnemolith

Mnemosyne — It provides sub-millisecond memory access with no external service costs.
claude-mem — It shares persistent, searchable agent memory across sessions and devices.
MemPalace — Achieves high recall with local-first storage and zero API calls.
Mnemoloci — It compiles projects into cited wiki pages and backs up vaults using git.
IHMT MEMORY — It shares one local memory across multiple AI agents without cloud storage or databases.
ProjectMem — Records project history locally to prevent AI from repeating failed work.
Mnemify — Runs locally on your machine and builds an interactive 3D knowledge map from multiple apps.
Engram — Single Go binary with zero dependencies and built-in full-text search.
AI Memory — Shared memory for coding agents that works across machines and team members.
Memorax Code — Carry project architecture and verified engineering decisions across coding sessions.
memsearch — It shares continuous context across Claude Code, Codex, and other AI coding agents.
AgentMemory — It runs locally with zero external databases and achieves high retrieval recall.

All Mnemolith alternatives