The problem with a general chatbot in a law school is not that it is unhelpful. It is that it will produce a citation that does not exist, in a format that looks exactly like one that does, and a second-year has no way to tell. LexForge retrieves the judgment first and will only discuss what it actually found — so a student can be marked on the output, and taught to check it.
A full academic term, every feature, every student, no card and no purchase order. We would rather find out whether your students use it than negotiate before they have touched it. If it turns out to be useful we will talk about what it costs; if it does not, it stops and nothing is owed. Student accounts and their work survive either way — ending a pilot never deletes anybody.
It does not write a memorial a student can submit unread. The draft is a structure and a starting position; the arguments still have to be theirs, and a judge will find out in thirty seconds if they are not.
It does not give legal advice, and it says so. What it produces is a draft for a qualified person to check.
It does not have every case. Live search covers reported judgments through Indian Kanoon and India Code; district court records and unreported orders are outside it. When a search finds nothing, it says nothing was found rather than filling the space.
Or just ask a question — whether it fits your syllabus, what the students actually see, whether it will survive a moot season.