Cross-Platform Semantic Memory System for AI Agents

memsearch is a cross-platform semantic memory system designed for AI coding agents. It allows memories to flow seamlessly across different platforms such as Claude Code, OpenClaw, OpenCode, and Codex CLI, enabling users to maintain a searchable context across conversations without additional setup.

Key features:

• Persistent Memory - Users can install a plugin to capture conversations automatically, providing semantic recall with minimal configuration.
• Markdown Files - Memories are stored as human-readable .md files, which are editable and version-controllable.
• Progressive Retrieval - The system employs a three-layer recall mechanism that includes search, expand, and transcript functionalities.
• Hybrid Search - Combines dense vector and BM25 sparse search methods for enhanced retrieval accuracy.
• Real-Time Indexing - A file watcher automatically indexes changes in memory files in real time.

This tool is particularly useful for both agent users, who benefit from easy installation and automatic memory capture, and for developers, who can utilize the CLI and Python API to integrate memory capabilities into their own agents.