procedural-memory

Human-like memory for AI agents — semantic, episodic & procedural. Experience-driven procedures that learn from failures. Free API, Python & JS SDKs, LangChain, CrewAI & OpenClaw integrations.

Kyros — The Memory OS for AI Agents Give your AI agents secure, self-correcting, persistent memory in 3 lines of code. Three memory types (episodic, semantic, procedural) with built-in forgetting curves, cryptographic integrity, and automatic contradiction resolution. Model-agnostic REST API with Python and TypeScript SDKs.

A dual-path memory system for proactive agents. Facts and procedures in a local knowledge graph, exposed as a CLI, MCP tools and skills. No API key.

Open-source procedural memory for AI agents. Learns from what worked and what failed.

Persistent filesystem-based memory system for Claude Code. Pure MIF Level 3 compliant memory storage with YAML frontmatter, bi-temporal tracking, and proactive hooks. No dependencies - just markdown files and git.

A persistent Minecraft inhabitant that remembers, feels functionally, forms habits, pursues goals, and learns how to live

MCP Context Server — a FastMCP-based server providing persistent multimodal context storage for LLM agents.

Memory for Claude Code that survives between sessions. Three kinds of memory (episodic, semantic, procedural) in one local markdown vault, surfaced in conversation through MCP.

LIA Labs / Learnkit - "Make your AI agents Learn Improve and Adapt".

A model-agnostic multi-layer vector storage system with explicit mappings and learned retrieval shortcuts.

Self-Improving Agents with Procedural Memory for LangChain. Agents that learn across task executions by extracting, storing, and retrieving reusable heuristics from their own experience. Research-backed feature inspired by HyperAgents, ERL, MARS, and DGM.

Governed procedural memory for AI agents — turn corrections, source material, and proven workflows into review-gated, versioned recipes agents can safely retrieve, lock, execute, and improve.

The audit and verification layer for AI agents - verified procedural memory with compliance reports (SOC 2, EU AI Act, NIST AI RMF), a public procedure registry, and the verifier gate

Procedural memory for LLM agents learn skills from failure, verify them before the agent trusts them

Agents should not relearn workflows. Myelin turns repeated agent behavior into reusable procedures via MCP.

Procedural memory for AI agents. Corrections become behavioral rules that compound over time.

Predecessor to Myelin. Broad local-first agent memory prototype; new procedural-memory work lives at github.com/Niraven/myelin.

A traceable reasoning architecture that converts experience into causally validated Reasoning KATA and reuses them with adaptive computational breathing.

Sleep-consolidated long-term memory for DeepSeek Harness (DSH) agents: episodic / semantic / prospective / procedural memory + identity, stdio JSON-RPC sidecar, idle-time dream consolidation, zero-config Cordis plugin. ｜ 给 DeepSeek Harness（DSH）agent 的睡眠整理式长期记忆：醒时回忆、睡时做梦整理，零配置 Cordis 插件。

A Python coding-agent harness for long-horizon work: procedural Skills, context folding, execution evidence and trace-grounded evaluation.