mempalace

Explore the MemPalace agent diary system and cross-wing navigation. Learn how this local-first architecture uses hierarchical drawers and graph tunnels for semantic concept connections across projects.

Deploy MemPalace with Docker and GPU acceleration. This guide details a multi-stage CUDA build for high-performance embedding inference using onnxruntime-gpu.

Explore MemPalace multi-language embedding support and configuration. Learn how to leverage the Embeddinggemma ONNX model and configure CPU fallback for efficient NLP.

Implement contradiction detection in MemPalace. This guide details offline comparison against local data for confusion and mismatch identification, no external APIs needed.

Master per-message verbatim recall with the MemPalace sweeper. Efficiently process Claude-Code JSONL transcripts for complete, resume-safe message storage.

Learn how to prevent prompt contamination with MemPalace query_sanitizer. Recover retrieval accuracy from 1% to 90% using our four-step strategy.

Learn input validation and sanitization in MemPalace config.py. Discover how strict functions protect data by stripping malformed Unicode, enforcing length limits, and validating dates.

Safely migrate MemPalace between ChromaDB versions. Our built-in tool rebuilds your palace extracting data from SQLite, ensuring no data loss. Learn more.

Easily repair MemPalace palace inconsistencies using the dedicated mempalace.repair module. Detect and fix HNSW index corruption with on-disk operations for enhanced safety and efficiency.

Learn MemPalace deduplication strategies for mined content. Discover how MemPalace uses a greedy longest match algorithm to remove redundant text and preserve complete documents.

Learn how to export palace data with MemPalace exporter module. Convert ChromaDB storage to Markdown safely and efficiently with export_palace().

Discover how MemPalace's local-first architecture ensures your data privacy by default. Learn the technical details of keeping information secure on your machine.