Agentmemory Architecture
rohitg00/agentmemoryCoding agents need a place to store what they learn. agentmemory is a memory server that runs on your own machine. It captures observations, indexes them, and serves them back over REST and MCP.
The system is built on the iii engine. Everything is a function, a trigger, or worker state. There is no separate plugin system. New capability means adding a function plus a trigger.
Recall uses hybrid retrieval with BM25 keyword search, vector similarity, and graph expansion. No API key is needed for basic use. An optional LLM provider adds richer summaries. A real-time viewer at port 3113 shows memory building as sessions run.
Add Agentmemory Architecture skill to your workflow
Global
mkdir -p ~/.claude/skills/agentmemory-architectureProject
mkdir -p .claude/skills/agentmemory-architectureSource Repository
More Memory & Knowledge Bases Claude Code Skills
Obsidian Baseskepano/obsidian-skills
Build powerful database views inside Obsidian with filters formulas and summaries
Compressjuliusbrussee/caveman
Shrink memory files into caveman speak to save tokens and preserve code
Remembergithub/awesome-copilot
Save lessons learned and build a smart knowledge base that grows with you
Remembering Conversationsobra/episodic-memory
Search past conversations to recall decisions, solutions, and avoid repeating mistakes
Claude Md Improveranthropics/claude-plugins-official
Audit your CLAUDE.md files for better AI agent project memory
Dbs Restoredontbesilent2025/dbskill
Continue your diagnosis where you last left off without repeating yourself
Winmd Api Searchgithub/awesome-copilot
Find Windows APIs easily with complete methods, properties, and events
Mem Searchthedotmack/claude-mem
Easily search past work across all sessions then filter before fetching details