Every contract tool produces a summary. A summary is exactly the wrong output, because the reader cannot tell the difference between a clause the model read carefully and a clause it skimmed, and the one that matters is always the one nobody read.
Briefly does not summarize. It flags, and every flag points at a character range in the original. Click it and you are looking at the sentence, in context, with the surrounding clause intact. If the model has nothing to point at, there is no flag.
What it looks for
- Uncapped or mutual-in-name-only indemnity.
- Liability caps that exclude the thing most likely to go wrong.
- Termination for convenience on one side only.
- IP assignment reaching past the scope of work.
- Payment terms that quietly restart on any dispute.