Tool
explore
Find anything. explore searches the index exhaustively — lexical and semantic together — and shows what exists, ranked by how well it answers you. One call reveals the landscape and the vocabulary: the real class names and terms-of-art for every question after it. Start here when you do not know what to read.
Four files named, ranked, and summarised for 620 tokens — and none of them opened. The north star document ranked above the implementation because the question asked where, not how.
Run it yourself
Or say it to your agent
“where is gitignore matching implemented in this codebase?”
Parameters
| Parameter | What it does | |
|---|---|---|
| keywords | required | Class names, concepts, or short phrases — the probe. No filler words. |
| question | optional | Your intent, as a full sentence. Results are reordered by how well they answer it, not just match it. Skip it only when surveying. |
| uriGlob | optional | Where to search. file:///** is the right default; unscoped spans every imported repository too. |
| breadth | optional | Maximum results. 2 for a deep dive, 10 for a survey, 0 to let the server choose. |
| tokenBudget | optional | How much room the answer gets — and therefore how deep it goes. |
Budget decides depth
The budget is not a cap on a fixed answer — it chooses the representation. Around 800 tokens maps what exists in headlines; 3,000 adds structure and snippets; 10,000 expands the key symbols in full. A cheap exploratory call that comes back wrong costs almost nothing; the read it saves you costs a great deal.
When results come back weak
Suspect your vocabulary before the repository. Terms borrowed from another codebase rarely land, and
keywords exists to map them to the real local names.
This page names what exists. The depth behind every name — the bounds, the failure modes, how they compose — ships inside the binary at help:///, and answers to explore and read exactly like your code does. Install it, and your agent has the manual.