fields.snippets to true. Each result’s snippets array holds its highest-scoring passages, in the order they appear in the document, so you can work with focused excerpts instead of a document’s full content.
Snippets are available on the
news and wikipedia scopes. The companies and people scopes don’t produce snippets; snippets comes back as an empty array for those scopes even when requested.Sample response
Get snippets alongside content
fields selects each member independently, so request content and snippets together to get both the full document and its top passages in one response.
Sample response
content here carries the article’s full 3,919 characters; snippets picked out a single sentence from a minister’s quote near the end of the piece, well past where this truncated content preview ends.
fields.content defaults to true, so omitting fields entirely returns content but not snippets. Once you pass a fields object, only the members you set to true are populated. Set content explicitly if you still want it alongside snippets.Tune how many snippets you get
Pass an object instead oftrue to bound how many snippets come back, and
how large each one is: max_snippets, max_snippets_per_result,
max_tokens, and max_tokens_per_result. max_snippets and max_tokens
bound the total across every document in the response; the _per_result
fields bound a single document. See the Search API
Reference for each field’s range and default.
See Errors for how an out-of-range value is rejected.
The example below caps each document at 2 snippets: