v3.0.0
MemPalace/mempalacev3.0.0Apr 6, 2026by bensig
AI Summary
Initial PyPI release - the highest-scoring AI memory system ever benchmarked at 96.6% LongMemEval R@5 (100% with Haiku rerank). Features the palace architecture with wings, rooms, halls, tunnels, closets, and drawers.
Key Highlights
- 96.6% LongMemEval R@5 - highest published score with zero API calls
- 100% LongMemEval R@5 with optional Haiku rerank
- Palace architecture - wings, rooms, halls, tunnels, closets, drawers
- AAAK compression - 30x lossless shorthand dialect for any LLM
- Knowledge graph with temporal entity-relationship triples in SQLite
- Specialist agents with per-agent wings and diaries
- MCP server with 19 tools
- Auto-save hooks for Claude Code
- Split command for transcript file handling
- Fully local - no API key, no cloud, no subscription
New Features
- mempalace split command for transcript splitting
- Specialist agent system with per-agent diaries
- Improved entity detection during init
- Room detection from folder structure
- Ruff-clean codebase
- AAAK compression dialect
- Knowledge graph in SQLite
- MCP server integration
- Auto-save hooks
Full Release Notes
## MemPalace v3.0.0 The highest-scoring AI memory system ever benchmarked. And it's free. ### Install ```bash pip install mempalace ``` **PyPI:** https://pypi.org/project/mempalace/ ### Highlights - **96.6% LongMemEval R@5** — highest published score with zero API calls - **100% LongMemEval R@5** — with optional Haiku rerank - **Palace architecture** — wings, rooms, halls, tunnels, closets, drawers — +34% retrieval from structure alone - **AAAK compression** — 30x lossless shorthand dialect, works with any LLM (Claude, GPT, Gemini, Llama, Mistral) - **Knowledge graph** — temporal entity-relationship triples in SQLite (local, free) - **Specialist agents** — each agent gets its own wing and diary in the palace - **MCP server** — 19 tools, auto-teaches AAAK to your AI - **Auto-save hooks** — for Claude Code (save every N messages + pre-compaction emergency save) - **Split command** — break concatenated mega-transcripts into per-session files before mining - **Fully local** — no API key, no cloud, no subscription. Everything on your machine. ### What's New in v3 - `mempalace split` command for splitting concatenated transcript files - Specialist agent system with per-agent diaries - Improved entity detection during `init` - Room detection from folder structure - Ruff-clean codebase ### Requirements - Python 3.9+ - No API key required