Integration

ScreenshotBase integration

ScreenshotBase renders web pages in a real browser: capture a URL as an image (png, jpg, webp, avif, gif, svg) of the viewport, the full page, an element or a region, print it as a PDF, render your own HTML or Markdown (invoices, reports, social cards) to an image or PDF, or extract a page's content as markdown. Slow pages can run as background jobs that you poll, download and delete, and you can read your quota and the emulated devices. The stored-URL upload mode, the render diagnostics output and the rendered-HTML download are not exposed. Authentication is an API key from the ScreenshotBase dashboard, sent in the apikey header; the free plan includes a monthly quota and 10 requests per minute. A successful render counts against the quota; most failed ones (a timeout, a blocked target) are not billed.

Connect ScreenshotBase How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in ScreenshotBase.Built from the 10 ScreenshotBase tools below, on the connection you add once.

What you can do with ScreenshotBase

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Job, 5 operations

create_render_jobCreate render jobPOSTStarts a render in the background and returns the queued job at once (up to 300 s, 120 on the free plan). Send exactly one of url, html and markdown. Poll get_job (or set webhook_url), then download the file with get_job_result. Failed jobs refunded.
delete_jobDelete jobDELETECancels and deletes a render job and its result. Queued: deleted and refunded. Rendering: answers 202, deleted and refunded when the render ends, result never delivered. A job that ran your scripts or an evaluate action keeps 1 unit. Free.
get_jobGet jobGETReturns a render job: its status (queued, processing, completed or failed), the result's URL, content type and size once completed, or the error the render answered once failed, and its quota charge. Free; poll it about once a second at most.
get_job_resultGet job resultGETDownloads the result of a completed render job as a stored file (the image or PDF the request without a job would have answered). Kept for 7 days after the job finished; a job that is not completed yet, or failed, answers 409. Free.
list_jobsList jobsGETLists your account's render jobs, newest first, a page at a time, optionally only those of one status. Jobs of a sandbox key are listed with a sandbox key only, the others with a live key. Free.

Screenshot, 2 operations

render_htmlRender htmlPOSTRenders HTML you send, or Markdown as a styled page, to a stored image or PDF file. Send exactly one of html (max 10 MB) or markdown (max 1 MB). Relative URLs do not resolve: use absolute, base or data: URLs. take_screenshot options; 110 s cap (408).
take_screenshotTake screenshotGETCaptures a public web page in a browser as a stored image (png default, jpg, gif, webp, avif, svg) or pdf (format): viewport, full_page, selector or clip. Over 110 s: 408, unbilled. HTML: render_html; text: get_page_markdown; slow: create_render_job.

Account, 1 operation

get_statusGet statusGETReturns your account's monthly and grace request quota (total, used, remaining) and its plan (name, billing interval, subscription status, renewal date, a scheduled plan change). Does not count against the quota or the rate limit.

Device, 1 operation

list_devicesList devicesGETLists the phones, tablets and laptops viewport_device accepts, each with its viewport size, device scale factor, mobile and touch emulation and user agent. Pass a device's id as viewport_device to take_screenshot. Free.

Page, 1 operation

get_page_markdownGet page markdownPOSTRenders a public web page (or HTML you send) in a real browser and returns it as markdown with final URL, title and meta description; main content only, ads and cookie banners blocked by default. Send exactly one of url and html. 110 s limit (408).

Use ScreenshotBase from Claude Code, Cursor or Codex

Connect ScreenshotBase in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its ScreenshotBase operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect ScreenshotBase

Create your workspace
Sign up free. Nothing to install.
Connect ScreenshotBase once
Add your ScreenshotBase connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.