> ## Documentation Index
> Fetch the complete documentation index at: https://exa.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> The Exa API is served at https://api.exa.ai. Authenticate with `Authorization: Bearer $EXA_API_KEY` (or `x-api-key: $EXA_API_KEY`); create keys at https://dashboard.exa.ai/api-keys.
> Prefer the official SDKs, `exa-py` (`pip install exa-py`) and `exa-js` (`npm install exa-js`); both read `EXA_API_KEY` from the environment.
> Tool-using agents can call Exa without writing code through the hosted MCP server at https://mcp.exa.ai/mcp, or install the Exa agent skill with `npx skills add exa-labs/agent-skills` (skill file: https://exa.ai/docs/skill.md).
> The OpenAPI specs at https://exa.ai/docs/exa-spec.yaml and https://exa.ai/docs/team-management-spec.yaml are the source of truth for request and response schemas.

# Research Publications

> Find academic papers, patents, grants, clinical trials, and regulatory approvals with Exa Search.

Use Exa Search for research publications and related records, including titles, abstracts, authors, venues, citations, publisher pages, preprints, and repository pages.

<Tip>
  Read [SOTA Search Over Academic Publications](https://exa.ai/blog/publications-search)
  to learn more about publication search quality.
</Tip>

## Included

* Papers and preprints, including full-text chunks where a parsed full text is available
* Patents, with abstracts, claims, inventors, and assignees
* Grants and funding announcements
* Clinical trials, drug labels, and interaction data
* Regulatory and health approvals

## Use it for

* Literature review and citation discovery
* Prior-art and patent landscaping
* Clinical and pharmaceutical research
* Grant and funding-opportunity discovery

## Example queries

### Find papers on a topic

Describe the method or finding rather than guessing at title keywords. The `publication` category keeps results to papers.

```text theme={null}
papers on evaluation benchmarks for retrieval-augmented generation
```

### Search clinical evidence

Name the phase, intervention, and population so trial registrations and results pages rank above general coverage.

```text theme={null}
phase 3 trials of GLP-1 agonists in adolescent patients
```

### Track regulatory approvals

Name the regulator and the device or drug class you are watching.

```text theme={null}
FDA approvals for AI-based diagnostic devices
```

### Run a prior-art search

Describe the invention functionally, the way a claim would, instead of using a product name.

```text theme={null}
patents on cooling battery packs with immersion dielectric fluid
```

## Make a request

<CodeGroup>
  ```python Python theme={null}
  from exa_py import Exa

  exa = Exa()

  results = exa.search(
      "papers on evaluation benchmarks for retrieval-augmented generation",
      type="auto",
      category="publication",
      num_results=10,
  )
  ```

  ```javascript JavaScript theme={null}
  import Exa from "exa-js";

  const exa = new Exa();

  const results = await exa.search(
    "papers on evaluation benchmarks for retrieval-augmented generation",
    {
      type: "auto",
      category: "publication",
      numResults: 10,
    }
  );
  ```

  ```bash cURL theme={null}
  curl -s -X POST https://api.exa.ai/search \
    -H "Authorization: Bearer $EXA_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "query": "papers on evaluation benchmarks for retrieval-augmented generation",
      "type": "auto",
      "category": "publication",
      "numResults": 10
    }'
  ```
</CodeGroup>

## Get structured data with Exa Agent

For structured data that requires research across multiple sources, use an [Exa Agent task run](/docs/agent/quickstart). Describe the publications, inclusion criteria, and output fields you need, and Agent returns schema-validated results with citations.

<Card title="Start an Agent task" icon="bot" href="/docs/agent/quickstart" cta="Open Agent guide" arrow="true">
  Build a literature map, screen papers against inclusion criteria, or gather fields from several publications into one table.
</Card>