소개
This tool empowers Large Language Models (LLMs) to effectively process and analyze massive amounts of text, exceeding 10 million tokens. Instead of feeding the entire context to the LLM, it implements the Recursive Language Model (RLM) pattern by intelligently chunking the data, performing recursive sub-queries on these smaller segments, and then aggregating the results for a final synthesis. It prioritizes cost-effective, free local inference via Ollama, falling back to cloud services like Claude SDK when local options are unavailable, ensuring efficient and scalable context handling for complex analytical tasks.
