Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
Why this repo matters
•Enables persistent agent memory across sessions via AI compression and context injection for multiple coding assistants.
•Supports diverse LLM backends including Claude, Gemini, and Copilot through a unified JavaScript memory engine.
•Integrates ChromaDB and SQLite for RAG-based long-term storage to solve stateless agent limitations.