Memory management for AI assistants. Automatic memory extraction and semantic retrieval for OpenWebUI, powered by Qdrant. Two components: an inlet/outlet filter for hands-free memory lifecycle and an
View on GitHub ↗A memory management system for AI assistants, built specifically for OpenWebUI. Bibliophylax gives your AI conversations persistent, searchable, semantically organized memory — with automatic lifecycle management and a clean MCP interface.
Bibliophylax is under active development. APIs, configuration options, and data models may change between releases. Contributions and feedback are welcome, but expect breaking changes until a stable v1.0.
Bibliophylax is a two-component system that provides AI assistants with persistent memory using a Qdrant vector database. It combines an automatic filter (inlet/outlet) that manages memory extraction and injection with an MCP server exposing on-demand memory tools — all backed by semantic search and structured lifecycle management.
OpenWebUI Filter (openwebui/filters/automemoryqdrant.py)
An inlet/outlet filter for OpenWebUI that manages the full memory lifecycle automatically:
- Inlet (before LLM call): Queries hot-tier memories semantically and injects relevant ones into the system prompt, capped by a configurable character limit. - Outlet (after LLM call): Extracts facts from the conversation via a structured LLM call, producing a MemoryActionPlan with add/update/delete actions. - Cold tier decay: Hot memories not accessed within a configurable period are automatically demoted to cold tier. - Cold→hot promotion: Relevant cold memories are promoted back to hot on query match. - Structured extraction: Uses Pydantic models for type-safe LLM output with deduplication and contradiction resolution.
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/gelse--bibliophylax.html)
Stop your AI from making things up — it proposes, deterministic tools decide, every claim checked against ground truth with evidence. Grounded facts and context survive resets. Reverse engineering is
zhuyansen/new-term-radarDaily radar for new AI terms: names of new models, products and repos, caught on X, Hugging Face, GitHub and sitemaps before keyword tools list them.
yzg1226/mcd-party-officer🍟 麦当劳门店派对成团调度 Skill — 把派对从「策划」推进到「办得成」：成团概率雷达 / 包场vs拼团 / 择日黄历 / 儿童营养指数 / 招募卡与邀请函
fellowgeek/mcp-memoryAn OKF-backed Model Context Protocol (MCP) server delivering persistent long-term memory and SQLite FTS5 search for AI agents.
hirotomasato/yowesGenerate realistic teacher documents (ID cards, licenses, letters) for 13 countries via MCP.
hujjun/mcd-saver-strategist麦当劳省钱点餐助手（麦门精算师）｜麦当劳优惠券组合优化 Skill：拉取麦当劳菜单与全部可用优惠券，算出最省的点餐组合，巨无霸套餐 ¥47 → ¥21.9（省 53%）。McDonalds MCP（麦当劳 MCP）coupon optimizer · McDonalds China MCP Server · Agent Skill for WorkBuddy · 麦当劳程序员创意开发大赛参赛作品
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.