Trimwise Prompt Compression

Description

Trimwise API enables efficient compression of lengthy documents into concise AI prompts while adhering to specified budgets. It supports both queryless and query-aware extraction methods, allowing users to select the most relevant excerpts from multiple sources under defined limits.

What can you use it for?

Build a retriever that keeps only the highest-value evidence for your prompts, compress multi-document context into a single budget, or create short, source-backed prompts for a chat model. You can trim by tokens, words, or characters and run it in queryless or query-aware mode. Try the hosted no-key endpoint for quick experiments or install the local Python library for private processing. It’s focused, extractive trimming — not an abstractive summarizer, so you keep original source spans.

API Stats