fast-rlm
What it is
A framework for Recursive Language Models (RLMs) that allows LLMs to interact with arbitrarily long prompts via an external REPL. It enables agents to write and execute code to decompose, filter, and transform data, returning results as symbols rather than expanding the context window.
The problem it solves
The limitation of LLM context windows when dealing with large datasets or long documents.
What you could use it for
- Processing and summarizing massive text files (e.g., millions of tokens) that exceed standard context windows.
- Automated data extraction and aggregation from large JSON/YAML files.
- Recursive task decomposition where sub-agents handle sub-tasks and return results as variables.
- Running local coding agents via ACP integration.
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.