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

# Exa

> Proxy neural web search, parsed page contents, and agentic answers through agntdata using your own Exa API key.

## Overview

[Exa](https://exa.ai) is a neural + keyword web search engine built for agents. The Exa connection on agntdata lets you proxy search, contents, find-similar, and agentic-answer calls with your agntdata API key — Exa bills you directly for usage.

## Prerequisites

* An agntdata workspace
* An Exa API key from [dashboard.exa.ai/api-keys](https://dashboard.exa.ai/api-keys)

## Connect

1. In the agntdata dashboard, open **Integrations** and click **Exa**.
2. Paste your Exa API key into the configuration dialog and save.
3. The key is encrypted at rest and only decrypted when proxying a request.

## Make requests

agntdata forwards the wildcard path and JSON body to `https://api.exa.ai` with your stored key attached as a bearer token:

### Example — neural search with contents

```bash theme={null}
curl -X POST "https://api.agntdata.dev/v1/connections/exa/search" \
  -H "Authorization: Bearer $AGNTDATA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "latest developments in LLMs",
    "numResults": 10,
    "contents": {
      "highlights": { "maxCharacters": 4000 },
      "summary": { "query": "Summarize key claims." }
    }
  }'
```

### Example — fetch parsed contents for URLs

```bash theme={null}
curl -X POST "https://api.agntdata.dev/v1/connections/exa/contents" \
  -H "Authorization: Bearer $AGNTDATA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "urls": ["https://arxiv.org/abs/2410.00001"],
    "text": { "maxCharacters": 8000 },
    "summary": { "query": "What does this paper contribute?" }
  }'
```

### Example — find similar pages

```bash theme={null}
curl -X POST "https://api.agntdata.dev/v1/connections/exa/findSimilar" \
  -H "Authorization: Bearer $AGNTDATA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://openai.com/blog/",
    "numResults": 10,
    "excludeSourceDomain": true
  }'
```

### Example — agentic answer with citations

```bash theme={null}
curl -X POST "https://api.agntdata.dev/v1/connections/exa/answer" \
  -H "Authorization: Bearer $AGNTDATA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "Who won the 2025 Nobel Prize in Physics and for what contribution?",
    "text": true
  }'
```

## Endpoints supported

* `POST /search` — neural + keyword search with optional contents enrichment
* `POST /contents` — parsed text, highlights, and LLM summaries for URLs or result IDs
* `POST /findSimilar` — pages semantically similar to a seed URL
* `POST /answer` — agentic Q\&A with cited sources (supports streaming)

## Rate limits

Exa applies per-key rate limits. agntdata surfaces upstream 429 responses as-is.

## FAQ

**Do I pay agntdata credits for Exa calls?** No. Exa bills your Exa account directly — agntdata only proxies the request and doesn't charge credits for `/v1/connections/*`.

**How is the API key transmitted to Exa?** agntdata injects your stored key as an `Authorization: Bearer` header. Exa also accepts `x-api-key`; either works.

**Does agntdata store my Exa API key?** Encrypted at rest in `workspace_integrations`; decrypted only when proxying a request.