AI Agents
agent-beacon
Self-improving memory layer that works across AI agent harnesses.
★ 1.8k starsMIT
View guide →A memory agent for your agent fleet: it curates, reconciles, and expires shared memories across tools.
memanto is an open-source ai agents project by moorcheh-ai (moorcheh-ai/memanto on GitHub). It has 2,320 GitHub stars and 735 forks, is released under the MIT license, and was last updated on GitHub on 7 October 2026.
Decide whether your agent memory should stay fully local (on-prem with Docker and Ollama) or use the free cloud backend.
Install Memanto from PyPI and launch it, choosing On-Prem or Cloud on first run.
Connect the coding agents you use, such as Claude Code, Cursor, or Codex, with the one-command connectors.
Store a first memory, recall it from a different agent, and optionally enable the scheduled daily curation loop.
Memanto extracts durable decisions, preferences, and facts from interaction streams instead of archiving whole transcripts.
Duplicates are merged, fragments joined, and contradictions supersede older beliefs while preserving what was believed and when.
Per-type retention tables and named rules expire stale memories on schedule; expired items stay recallable, clearly labelled, until you delete them.
Memories export in the Open Knowledge Format as plain Markdown, and migrate imports from tools such as Mem0, Letta, Supermemory, Zep, and Hindsight.
Thirteen memory categories (instruction, fact, decision, preference, and more) plus point-in-time and changed-since recall queries.
Agents are briefed with the minimal relevant slice before acting, and a local UI plus daily summary show the state of the whole estate.
Most users install from PyPI; the source lives here.
git clone https://github.com/moorcheh-ai/memanto.git
cd memantoInstall the CLI, run it once to pick On-Prem or Cloud, and connect your agent.
pip install memanto
memanto
memanto connect claude-codeStore a typed memory in one agent and recall it from anywhere in the fleet.
memanto remember "Auth migrated to JWT" --type decision
memanto recall "how does auth work"It is maintained by moorcheh-ai (Moorcheh) on GitHub.
Yes. Memanto is MIT-licensed with no open-core gating; on-prem mode is free, and the cloud option offers a free key with a free operations allowance per the README.
Not in on-prem mode: it runs locally with Docker and Ollama and needs no account or API key. Cloud mode uses a free Memanto API key rather than your LLM key.
Tools like Mem0 or Zep primarily store and retrieve memory; Memanto positions itself as an agent that actively curates, reconciles, expires, and briefs, and it can import from those tools.
Run pip install memanto, launch memanto, choose On-Prem or Cloud, then connect your coding agent with memanto connect.
Self-improving memory layer that works across AI agent harnesses.
A benchmark for comparing AI agent memory systems head-to-head.
Lean coding agent written in Rust, optimized for memory footprint and performance.
30 runnable notebooks teaching everything about agent memory.
Monid — OpenRouter for agent tools: one base URL and one key give an agent 2,000+ tools across 72+ providers
Agent memory that learns — so your agents get smarter over time.
The most-starred other AI Agents projects in the LibreAI catalogue — the realistic shortlist if memanto is not the right fit.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Free and open — grab the source, star the repo, and start creating.