Memento-Teams/Memento
A memory-based continual-learning framework that improves LLM agents through experience without updating model weights.
Not currently ranked — collecting fresh signals.
star history
Memento implements a planner-executor architecture with case-based reasoning for LLM agents. The system maintains a memory store that accumulates agent experiences and retrieves relevant past cases to inform current decisions, enabling agents to learn from their history without modifying underlying model parameters. It supports MCP tool integration and vLLM for local LLM serving, with evaluations conducted on the GAIA benchmark.
Frequently asked
- What is Memento-Teams/Memento?
- A memory-based continual-learning framework that improves LLM agents through experience without updating model weights.
- Is Memento open source?
- Yes — Memento-Teams/Memento is open source, released under the MIT license.
- What language is Memento written in?
- Memento-Teams/Memento is primarily written in Python.
- How popular is Memento?
- Memento-Teams/Memento has 2.5k stars on GitHub.
- Where can I find Memento?
- Memento-Teams/Memento is on GitHub at https://github.com/Memento-Teams/Memento.