umerkay bolts a web research desk onto Mike

Four new endpoints let the backend scrape pages, ask questions of a URL, crawl whole sites, and map their structure.

searchintegration

umerkay has wired Mike up to Olostep, a third-party web-research service, giving the backend the ability to pull a clean text version of any page, ask an AI questions about what's on it, crawl across multiple pages, or map out the shape of a site. The plumbing is opt-in: if you don't set an API key, nothing breaks, and nothing costs you anything.

The target use case is legal research - fetching cited sources, interrogating a page for the answer to a specific question, or building a local picture of a publication's site. It's a clean bolt-on rather than a deep integration: no frontend yet, no rate limits or audit logging on the outbound paid calls, and the crawl endpoint hands back a job ID with no way to retrieve the result in this cut. The shape of the integration also ties Mike fairly tightly to Olostep specifically.

So what Worth a look for any team thinking about giving their legal-AI tooling a real research-the-open-web capability, with the caveat that the guardrails aren't in the box yet.

View this fork on GitHub →

Spotted something wrong? Or know the PR text has fresher detail than the writeup above?

Commits in this thread

1 commit from umerkay/mike, oldest first. Source extracted verbatim from the harvested git log.

SHA Subject Author Date
007f8123 feat(research): add Olostep web research endpoints umerkay 2026-05-01 ↗ GitHub

Capture this thread into my fork

Download a single Markdown prompt that tells Claude how to port every commit above into your working tree — adapting paths and structure to match your repo. Run it via claude -p < capture-thread-140.md from inside the repo you want the changes in.

⬇ Download capture-thread-140.md