Search the Neutrinos product documentation (53 publications, 3,117 topics) and return ranked, citable passages. This is the entry point for any question about Neutrinos products. IMPORTANT: the same topic is often documented separately for several product versions. If the user's version is known, pass product and version; otherwise only current versions are searched. Check sufficient_evidence before answering - when it is false, say the documentation does not cover the question rather than inferring an answer. Call list_products first if unsure what to pass for product.

Retrieve the full text of a documentation page, or one section of it, by the ref returned from search_docs. Use this when a search passage is clearly relevant but cut off mid-explanation or mid-code-block. Prefer passing section - whole pages can be long, and the response is hard-capped at max_tokens.

Return the typed neighbourhood of a documentation page: its parent and children, the previous and next pages in authored reading order, curated 'see also' links, in-prose cross-references, and the same topic in other product versions. Use this when a page assumes a prerequisite you have not read, or when you need to confirm which product versions document a behaviour. prev is the best signal for 'what should the user have set up first'.

Show how one documentation topic differs between product versions, section by section. Use this for any 'this worked in version N but not N+1' question, or before answering when search_docs set version_ambiguous. Note that some products were renamed between versions (App Builder became Studio, for example) - this tool follows renames, so it spans them where a name-based search would not.

List every Neutrinos product in the documentation with its versions, which version is current, how many topics it has, and when it was last updated. Call this before search_docs when you are unsure what to pass for product or version - guessing filter values is the most common cause of an empty search. Cheap and cacheable; the answer only changes when the index is rebuilt.

Explores thematic relationships between entities, returning nodes and edges up to max_depth hops. Use this to understand structural or dependency links between concepts rather than text search.