> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.usescout.sh/llms.txt.
> For full documentation content, see https://docs.usescout.sh/llms-full.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.usescout.sh/_mcp/server.

# Introduction

Scout turns the open web into clean, structured data your application can use
directly. It is a single REST API: send JSON, get JSON. There is no SDK to
learn and no scraping infrastructure to run - every endpoint returns results
already shaped for large language models and autonomous agents.

## The API

All endpoints speak JSON over HTTPS and live under one base URL, versioned
under the `/v1` prefix.

```bash Base URL
https://core.usescout.sh
```

## Capabilities

Seven endpoints cover the common web-intelligence workloads.

Rank the live web for a query; results carry title, URL and excerpts.

Turn public URLs into clean, LLM-ready markdown content.

Deep research - a multi-step agent that answers with a per-field citation basis.

An OpenAI-compatible chat endpoint grounded in live web research.

Enumerate every entity matching a description as structured rows.

Track a query on a schedule; get webhooked when results change.

## Your first request

Create an API key in the [Scout dashboard](https://platform.usescout.sh) under
**Settings → API keys**, then pass it as a Bearer token to any `/v1` endpoint.

```bash cURL
curl -X POST https://core.usescout.sh/v1/search \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"queries": ["what is retrieval augmented generation"], "limit": 3}'
```

Results are sourced from the live public web and reshaped for agent
consumption - just the URL, title and excerpt your model needs to reason and
cite. Continue to the [Quickstart](/quickstart) to see the response shape and
the [Authentication](/authentication) page for key management.