15 building blocks of AI agent memory, distilled from building 400+ agents
MaryamMiradi · x · 2026-11-14
AI engineer Maryam Miradi distills 15 memory building blocks for agents from 400+ agent builds, with open-source GitHub links, organized into three groups:
Memory Types (what to remember)
- Working, episodic, semantic, procedural, and long-term memory
Memory Mechanisms (how to process)
- Summarization, retrieval, consolidation, compression, graph memory
Memory Control (what to trust/forget)
- Decay, forgetting, conflict resolution, provenance tracking, memory isolation
The framework separates what agents remember, how memories are processed, and what gets trusted, forgotten, or isolated.
More from coding & agent
- FastMCP 4.2 Ships Skills Over MCP, Pairing Tools With Usage Guidance — jlowin123 · 2026-11-14
- Dev reverse-engineers Waymo's private API into a Cloudflare Worker MCP for Claude — MicahBerkley · 2026-11-14
- Dev's fix for chaotic agent work: a feed-shaped todo list with action buttons — _rchaves_ · 2026-11-14
- Dev replaces his smartphone with 8 Raspberry Pis controlled by OpenAI's Dot — OpenAIDevs · 2026-11-14
- Garry Tan open-sources idea-miner skill that turns your last 48 hours into a ranked build list — garrytan · 2026-11-14
- LLMLean is back: LLM-powered proof suggestions for Lean, local or cloud — wellecks · 2026-11-14