wardethan2000-eng makes legal research runs harder to lose
This fork is turning long AI research sessions into a more dependable route to usable legal sources.
When a research run reaches its limit, it now returns a grounded partial answer instead of stopping silently. Users can then continue the same thread with a Keep going prompt across the major AI providers supported by the product.
The legal-research side is becoming more concrete:
- Numbered statute sources that can be opened in a panel for closer reading.
- A dedicated statute service, aimed at making legislation a sustained source for research rather than a one-off chat reference.
- Document-editing improvements bundled into the same broad integration.
This is a substantial, cross-provider change rather than a small plug-in. Anyone bringing it into another installation should inspect how limits, continuation state, source-service access, and provider testing fit their own setup.
Spotted something wrong? Or know the PR text has fresher detail than the writeup above?