Corey Web and Research Tools: Browse, Search, Read

Web and research tools

Corey can browse the web, run searches and read pages to gather what a job needs - market research, references, current facts. It works with the open web on your behalf, and is clear about what it can and cannot reach, such as sites that sit behind a login.

A lot of work starts with finding things out. Web and research tools, reached over MCP, let Corey browse, search and read pages so it can gather the facts, references and context a job needs - market research, current numbers, prior art - before it drafts a word.

This is research on your behalf against the open web. Corey is clear about what it can and cannot reach, and it brings back sources, not just claims.

What Corey can do on the web

For a research job, Corey can:

search the web for sources relevant to the task
open and read pages, then pull out what matters
summarise findings into something you can act on
cite what it used, so you can check the source

It treats the web as an input to the work, not the work itself - the point is the plan, draft or decision it informs.

What it will not reach

Corey works with the open web. It does not quietly break into places it should not be:

pages behind a login or account wall
paywalled or subscription-only content
anything that needs credentials you have not connected

When something is out of reach, Corey says so plainly rather than inventing an answer.

How research is captured

For a quick lookup, Corey answers in the conversation. For anything substantial, it writes the research up as a file in your Corey folder - a summary with the sources behind it - so the work is something you keep and can reuse, not a message you have to find again later. See file and folder tools for how that output is stored.

Reading freely, acting carefully

Browsing and reading are read-only, so they run without interruption - the same approval model that runs across every tool. The moment research turns into something that goes out - an email, a published page - that step waits for your approval. See what you can ask Corey for the kinds of research jobs people hand over.

Questions, answered

Yes. Corey can run searches and open pages on the open web to gather current facts, references and context. It is reading the real web, not a fixed snapshot, so it can pull in recent information a job needs.

Not by default. Corey works with the open web. Content that sits behind a login, a paywall or a consent wall is out of reach unless you connect a tool that provides that access, and it will tell you when it cannot reach something rather than guess.

Either back into the conversation or saved as a file in your Corey folder, depending on the job. For anything substantial - market research, a competitor scan - Corey writes it up as a document you keep, with the sources it used.