umerkay wants Mike to reach out to the open web
A proposed integration would let the assistant pull in live content from across the internet, not just the files a user uploads.
umerkay drafted a way to plug Mike into Olostep, a hosted service that fetches and reads web pages, and exposed it as four research tools:
- Grab a single web page on demand and pull its content into the assistant.
- Ask a question and get an answer drawn from live web content, not just uploaded documents.
- Crawl an entire site to gather everything behind it.
- Build a map of a site's pages so the assistant knows what's there before it digs in.
The whole thing was designed to sit quietly alongside the existing product: if the outside service isn't set up, nothing else breaks. But the proposal was closed without being merged, so none of it is live today. It stands as a blueprint for how Mike could break out of its own document library and start researching the open web.
So what Anyone weighing whether a legal assistant should stay walled inside uploaded files or reach live sources should read this as a preview of the second path.
Spotted something wrong? Or know the PR text has fresher detail than the writeup above?