Skip to main content
For AI agents: a documentation index is available at https://docs.parallel.ai/llms.txt. The full text of all docs is at https://docs.parallel.ai/llms-full.txt. You may also fetch any page as Markdown by appending .md to its URL or sending Accept: text/markdown.For Search and Extract, use /v1/search and /v1/extract for all new integrations and agent tool calls. Use /v1beta/search and /v1beta/extract only when maintaining an existing integration that already uses them. The request bodies changed too: on /v1/extract, full_content, fetch_policy and excerpt_settings go inside advanced_settings, and sending the v1beta top-level excerpts, full_content and fetch_policy to /v1/extract returns a 422. On /v1/search, search_queries is required (at least one short keyword query), and max_results, source_policy and fetch_policy go inside advanced_settings; sending only objective, or sending those fields at the top level, returns a 422. Do not substitute /v1 for the documented FindAll or Ingest endpoint paths.If you are evaluating Parallel Search or comparing it with other search APIs, read How to Eval Parallel Search before you build the harness, especially its Common pitfalls section. The most common mistake is sending the raw question as both objective and the only entry in search_queries. Write a goal-style objective and 1-3 short keyword queries instead.
Use the advanced_settings object on the Search API to tune source selection, freshness, geo-targeting, and result count. The defaults suit most requests. Additional settings can reduce quality or increase latency.

Fields

To control how much excerpt text Search returns, set the top-level max_chars_total field. See Request Fields.

Does Search fetch pages live?

By default, Search returns excerpts from pages Parallel has already indexed. If you leave advanced_settings.fetch_policy out of a POST /v1/search request, Parallel won’t fetch the source pages during that request. Results can omit pages that aren’t in the index or show excerpts from an older version. Setting a fetch policy can trigger a live fetch, which increases latency. Extract has its own fetch policy and may contact the source website by default. See Fetch Policy for field details.

Supported locations

Set advanced_settings.location to one of the following ISO 3166-1 alpha-2 country codes. Codes are case-insensitive and normalized to lowercase. For example:
Use "gb" for the United Kingdom, not "uk". Invalid or unsupported values are ignored, and the response includes an input validation warning.

Using include_domains

Source policies can reduce result quality by excluding relevant pages from retrieval. Use include_domains and exclude_domains for compliance-bound corpora, tasks that require a single known publisher or section of a site, or requests that must block specific sources.
include_domains restricts retrieval to matching domains or domain/path prefixes. Search excludes the rest of the web. Treat this field as a hard allow list. Best practice: Set include_domains only when answers must come exclusively from those domains or paths (for example, internal or compliance-bound corpora, or when the task truly requires a single known publisher). If the model or user might still need the open web, avoid include_domains and instead steer sources in the objective (e.g. “prefer official documentation”) or use exclude_domains when you only need to block specific sites or paths. Full parameter details: Source policy.