Skip to Content
Developer’s Eye

Feature overview

What you can do with the search interface and what has to be running behind it.

Find a specific page

Start with the name of a library, a concept or a symbol. All searches across prose and code. Text is handy for a question like “how does async work?” and Code is handy when you remember an identifier. Title matching also helps surface a reference page whose heading matches the query.

Source filters narrow the set to categories such as documentation or GitHub. The categories present in a live response depend on what your crawler has actually indexed. A matching passage under the title helps you decide whether to open a result. Code previews expose examples, and saved results stay in that browser.

Recover from a typo

When the index has enough evidence, search can offer a “did you mean” link. Accepting it starts another search. The original query and its results remain available until you choose the suggestion. The candidate words come from the indexed corpus, so an empty or freshly built index can offer few corrections.

Read around a result

The knowledge panel looks up a topic on Wikipedia. It appears when the title or search match is confident enough. The optional AI overview works from evidence in the returned results and cites its sources. Each feature has its own request path, so ranked links can still be useful when an enrichment is unavailable.

Watch the engine work

Operators can see which hosts the crawler is visiting, whether indexing is keeping up, and how the corpus is growing. The search-quality tools let a reviewer grade results for a query. Follow Admin and observability to connect the dashboard.