May or Shall

Legal AI has learned to summarise, but I am not sure it has learned how lawyers think. Most legal AI tools make summarising, retrieving and drafting faster, whereas the real craft of law lies in judgment, i.e. reading, identifying what matters, spotting contradictions and connecting facts to the argument they shape.

With that idea, I built "May or Shall". It is a Chrome extension, a website, a desktop app for Windows and Mac and an app inside ChatGPT, which allows a person to read, highlight the lines that matter and tag them under a heading, with every line keeping its source, i.e. the page and the paragraph it came from. Those lines can then be linked across documents, so that a paragraph of a plaint can be connected to the exhibit that proves it, and ChatGPT drafts from them through the May or Shall app.

The human in the loop can also be at the start of the task instead of just the end. The judgment stays with the person who read the document and only the typing moves, and every sentence of the draft can be checked against the line it came from. I am a lawyer and I built it with Claude Code.

• Save the lines that matter as you read, from any web page, PDF or app, each keeping its page and paragraph
• Tag them under legal headings, such as Fact, Issue, Admission and Date, or under headings of your own
• Link a line in one document to the line it relates to in another, e.g. a plaint paragraph to the exhibit that proves it
• Draft with ChatGPT from your own highlighted lines through the May or Shall app in ChatGPT, with every source a click away
• A list of dates built from the dates you mark, exported to Word
• Works as a Chrome extension, a website and a desktop app for Windows and Mac, which saves from WhatsApp, Word or Outlook also
• Catch the AI, a game on Shreya Singhal v. Union of India in which you highlight the words that prove an AI summary wrong

I would value feedback from lawyers and law students who read a lot before they draft, in particular on where reading and highlighting in May or Shall feels slower than it should, and introductions to law schools and firms which would like to try it with their teams.