Hi everyone! I wanted to share a Python script I recently developed to help manage large text inputs for LLM context windows. The script compresses lengthy files while preserving key information, using one of six compression techniques (e.g., bullet points, key points, paraphrasing). This lets you fit essential content within token limits, freeing up space for more instructions or context without risking truncation of long documents.
If you’re working with large reference texts or RAG-like setups, this might be helpful! Here’s the GitHub repo if you’d like to check it out or provide feedback: