In short. You.com Research API combines web research with referenced answers and controls for sources, research effort, structured output and background execution.

Definition: You.com Research API is a programming interface that searches the web, reads sources and synthesizes an answer with references. As of 17 October 2026, You.com documents Markdown answers with inline citations and an optional structured JSON response.

Key facts

You.com's Research API reference specifies POST https://api.you.com/v1/research, authenticated through the X-API-Key header.
The Research API Quickstart lists the research_effort settings lite, standard, deep, exhaustive and frontier; standard is the default.
The Source Control guide documents domain lists limited to 500 entries, including an allowlist and an exclusion list.
The Structured Output guide states that sources remain in output.sources when the answer uses a JSON schema.

What does a sourced response contain?

The Research API Overview describes a process that performs multiple searches and cross-references the retrieved material. Its default answer is Markdown text in output.content. An inline reference such as [[1, 2]] points to entries in the accompanying sources array.

The API reference specifies a url for each source record, with optional title and snippets fields. Snippets contain relevant excerpts used to generate the answer. For a website review, these records provide the source pages and supporting extracts to examine alongside the answer.

How does source selection work?

The Source Control guide defines source_control as a request field. Within it, include_domains restricts results to named domains. exclude_domains removes named domains from results and blocks the research agent from visiting their pages during browsing.

You.com's announcement about research controls specifies that those lists are mutually exclusive within a request. It also documents freshness values such as day, week, month and year, while country focuses results geographically using country codes such as GB.

This illustrative request body follows the guide's domain-filtering example. The question text is a placeholder:

{
  "input": "<research question>",
  "research_effort": "deep",
  "source_control": {
    "include_domains": ["arxiv.org"]
  }
}

How does structured output handle references?

The Structured Output guide documents output_schema, a JSON Schema definition whose root must be an object. It changes the answer in output.content into an object; output.content_type becomes object.

The source list remains outside that object in output.sources. The API adds no citation fields automatically inside the requested schema. An application displaying structured findings therefore needs to handle the source list alongside those findings.

How are depth and background execution controlled?

The Quickstart presents lite for quick lookups and deep or exhaustive for tasks where thoroughness takes priority over speed. It reserves frontier for longer research tasks requiring background requests.

The Background Requests guide documents synchronous execution by default. Setting background: true returns a task handle immediately. The application checks progress through GET /v1/research/{task_id} and retrieves the completed result separately from submission.

FAQ

Can I restrict research to selected websites?

Yes. The Source Control guide documents include_domains for this purpose, with up to 500 domains. An exclusion list also blocks browsing of the excluded domains.

Where can I inspect the supporting material?

Use output.sources. The API reference specifies source URLs and optional titles and snippets, so you can open the referenced pages and compare their contents with the answer.

Sources

Check how a page reads to answer engines. Open the GEO score analyzer