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

# ScreenshotAPI documentation

> Turn web pages into images and PDFs, then build captures into your application.

ScreenshotAPI renders a URL in a browser and returns a downloadable capture with metadata. Use it for website previews, reporting, visual monitoring, or tools that need to inspect a web page.

You choose the viewport, output format, page readiness rules, and content controls. The REST API handles queued execution and returns either the finished capture or a job you can track.

<CardGroup cols={2}>
  <Card title="Make your first capture" icon="bolt" href="/quickstart">Create an API key, submit a URL, and download the result.</Card>
  <Card title="Understand the response" icon="file-lines" href="/responses">Handle finished captures, accepted jobs, and structured errors.</Card>
  <Card title="Capture a full page" icon="expand" href="/capture/full-page">Configure scrolling, resolution, and capture diagnostics.</Card>
  <Card title="Run captures in the background" icon="clock" href="/integration/async-jobs">Track queued work without holding an HTTP connection open.</Card>
</CardGroup>

## Choose your integration

| Integration | Use it when |
| - | - |
| REST API | Your application needs hosted rendering, job tracking, and account billing. |
| [Node.js](/integration/node) or [Python](/integration/python) | You want a complete example using standard HTTP libraries. |
| [Hosted MCP](/agents/mcp) | An AI agent needs authorized captures and page inspection using OAuth or an API key. |

The hosted REST API base URL is `https://api.screenshotapi.tech`. The examples include this address directly; only `SCREENSHOTAPI_API_KEY` needs your private value. The documentation domain is a separate site and does not accept capture requests.

## What to expect

* Image formats are PNG, JPEG, and WebP. PDF captures preserve the rendered appearance of the page.
* A successful capture includes dimensions, byte size, final URL, timing, and an artifact URL.
* A request can remain queued after the HTTP response. Treat `202 Accepted` as a job handle and wait for its terminal result.
* Failed captures release their reserved credits. Credit cost depends on the requested output.
* Public API captures require public network destinations. Target access restrictions and bot challenges can still prevent a capture.

Continue with the [quickstart](/quickstart), then add [idempotency](/integration/idempotency) and [webhooks](/integration/webhooks) when your workflow needs reliable background execution.

For AI workflows, connect the [hosted MCP service](/agents/mcp) or request [page text with your capture](/capture/inspection). You can grant a dedicated integration only the permissions it needs.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.