agentmemory

Recursive Experiential–Working Memory Evolution for Long-Horizon Agent Harnesses

Long-term, cross-project memory for AI coding agents. Your own Obsidian vault as the source of truth. Daemonless and without opaque databases, your memory belongs to you.

Neuroscience-inspired memory framework for AI agents

Agent Memory Playground: AI Agent Memory Design & Optimization Techniques

Pull-model episodic memory plugin for Hermes Agent. Real deletes, audit trace, BYO Claude. MIT.

Implementation based on the paper "ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory"

Aurora is a first-principles adaptive memory runtime organized around one evolving trace field.

Your agent, do things in your way.

A toy local memory system for multi-agent workflows

Provider-free persistent cognition for AI agents: verified memory, failure frontiers, proof-gated branches, and reversible learning.

A dual-layer memory system for AI agents with incremental clustering, graph-based retrieval, and conflict resolution.

A production-grade memory system for AI agents that enables accurate recall across 1,000+ conversation turns.

HTM-EAR: A hierarchical, importance-aware memory architecture for long-running LLM agents, combining tiered vector storage, entity-aware retrieval, policy-driven retention, cross-encoder reranking, and saturation-stress evaluation. Research implementation accompanying arXiv:2603.10032.

A minimalist, cross-platform personal AI automation toolkit for research, learning, and lead generation. Built for Termux + WSL2 + Windows with Gemini CLI, AgentMemory, Apify, yt-dlp, and gcloud.

agentmemory Viewer 简体中文汉化补丁 — 纯文案替换，不改功能逻辑、不发送数据，覆盖 295 条界面文案（适配 agentmemory v0.9.29）

Voice agent system for learning a new language

The first collision-based memory system: it doesn't retrieve old ideas — it smashes old cards into new ones.

🛠️ Experiment with AI agents in a local sandbox to run, score, and compare outputs without the need for complex frameworks.

MCP stdio → Streamable HTTP (Python/Starlette): `agentmemory` serves as shared long-term memory for Freebuff and OpenCode — includes a watchdog, `/health` status (ok|degraded|down), systemd units, and E2E testing.