Name: ScreenshotAPI
Author: Screenshot-API

ScreenshotAPI MCP

Website screenshots, URL/HTML to PDF, scrolling videos, and web scraping via ScreenshotAPI.

Grade A Local · stdio API key 8 tools v1.0.2 MIT

No reviews yet by Screenshot-APInpm 243 Runs on your machine, nothing hosted

96/100

trust score

downloads a day

local

runs on your machine

243

downloads in 30d

What it does

Website screenshots, URL/HTML to PDF, scrolling videos, and web scraping via ScreenshotAPI.

Quickstart

# 1 — add it to Claude Code
claude mcp add --scope user --env 'SCREENSHOTAPI_TOKEN=${SCREENSHOTAPI_TOKEN}' --transport stdio screenshotapi-mcp '--' npx -y @screenshotapinet/mcp

# 2 — ask your agent for something these tools do
#   take_screenshot, generate_pdf, scrape_page and 5 more

It needs SCREENSHOTAPI_TOKEN: this reads it from your environment, so export it first.

ScreenshotAPI costs nothing on mcprush: there is no plan to choose here, no cap to set and nothing mcprush can bill you for.

Publishernot claimed

Screenshot-API1 server

1 category

All 1 server

Collected from a public index. Nobody has claimed this account, so nothing here was written by its author — claim it if it is yours.

Installation

Where are you running it?

Every route below installs the same thing and ends at the same approval screen. This one runs on your machine: your client starts ScreenshotAPI as a process under your own user, with your files and your network, so the tool surface below is what it can do to your computer rather than to a server somewhere else. Its scan result is on the Security tab — read the surface before you approve it.

Claude Codedeclared by this listing

This is a public server: you run it yourself and this marketplace is not in the path. You need Node.js — the current LTS (24) is the safe choice; npx comes with it. Claude Code registers it in one command; --scope user makes it available in every project. Claude Code reads SCREENSHOTAPI_TOKEN (Your ScreenshotAPI API key from https://app.screenshotapi.net) from your environment each time it starts the server, so the value stays out of its config: export it in your shell profile (macOS, Linux), or run setx SCREENSHOTAPI_TOKEN … and open a new terminal (Windows). The '--' is in quotes so that PowerShell passes it on.

claude mcp add --scope user --env 'SCREENSHOTAPI_TOKEN=${SCREENSHOTAPI_TOKEN}' --transport stdio screenshotapi-mcp '--' npx -y @screenshotapinet/mcp

Reconnect, or start a new session, and the tools appear in the model’s tool list.

Free on mcprush, and it runs from its publisher’s own package: this marketplace is not in the path, so there is no key of ours to hold and no card to add.

What it asks of you

It reaches a system of yours, so it needs your own credential rather than the publisher’s: SCREENSHOTAPI_TOKEN — Your ScreenshotAPI API key from https://app.screenshotapi.net. You set it in your own client’s config, on your machine — this marketplace never holds it.

What it writes

One config entry your client uses to start the process locally. A local server runs with your file system and your network, which is why it is priced without metering.

What you approve

8 tools, with what each one reads, writes and reaches shown before you agree — the same list on every route above. Read the tool surface.

Tool surface

What the model actually sees. Descriptions are diffed on every release — see version history.

Only tools that write or reach the network

Read from its source code. A server that builds its tool list when it starts can list more tools than this.

Captures a screenshot of a web page (or of raw HTML) in a real browser and returns the hosted image URL. By default the image is also returned inline so you can see it — except for full_page captures, which are often too tall to view; set include_image=true to force it.

Parameter	Type	Required	Notes
file_type	string	no	Image format. Default png. jpg is smallest (see image_quality); webp balances size and quality. For PDF use generate_pdf.
full_page	boolean	no	Capture the entire scrollable page instead of just the viewport. Default false. Combine with lazy_load for pages with lazy images.
retina	boolean	no	Render at 2x pixel density for a sharper, larger image. Default false.
image_quality	integer	no	JPEG compression quality 0-100 (jpg only). Default 80.
omit_background	boolean	no	Make plain white/solid page backgrounds transparent (png only). Default false.
thumbnail_width	integer	no	Resize the output to this width in pixels, keeping the aspect ratio. Handy for previews and to keep inline images small.
selector	string	no	CSS selector of a single element to capture instead of the viewport, e.g. '#pricing' or '.product-card'.
clip	object	no	Capture only this rectangle of the page.
sizes	array	no	Capture the page at several viewport sizes in one request, e.g. [{width:1920,height:1080},{width:768,height:1024},{width:390,height:844}]. Overrides width/height. Returns one hosted image per size; im
include_image	boolean	no	Return the image inline in the tool result so you can see it. Default true, or false when full_page is true. Images above SCREENSHOTAPI_MAX_INLINE_IMAGE_BYTES (default 3500000 bytes) are returned as a
save_path	string	no	Optional local path to also save the file to, e.g. './shots/home.png'. A directory path (ending in '/' or an existing directory) keeps the remote file name. Relative paths resolve against the MCP serv

Renders a URL or raw HTML in a real browser and converts it to a PDF, returning the hosted PDF URL. Choose a paper size ('format') for paginated output; without format (and without width/height) the whole page becomes one long continuous PDF page.

Parameter	Type	Required	Notes
format	string	no	Paper size. A4 or Letter for documents and invoices. Omit for a single continuous page.
landscape	boolean	no	Landscape orientation — good for wide tables and dashboards. Default false (portrait).
page_range	string	no	Keep only these pages, as 'start-end', e.g. '1-3' or '5-5'. Needs a paginated PDF (format or width/height).
print_background	boolean	no	Include background colors and images in the PDF.
media	string	no	CSS media type to emulate: 'print' uses the site's print stylesheet, 'screen' matches what a visitor sees.
save_path	string	no	Optional local path to also save the file to, e.g. './shots/home.png'. A directory path (ending in '/' or an existing directory) keeps the remote file name. Relative paths resolve against the MCP serv

Loads a page in a real browser (JavaScript executed) and extracts its content: clean Markdown (best for reading and summarizing), plain text, the fully rendered HTML, and/or every image URL on the page. Extracted files are fetched and returned inline, truncated to max_content_chars.

Parameter	Type	Required	Notes
extract	array	no	What to extract. Default ['markdown']. Options: markdown, text, html, image_urls.
include_content	boolean	no	Fetch the extracted files and return their contents inline. Default true. false returns only the file URLs.
max_content_chars	integer	no	Truncate each extracted file to this many characters. Default 50000.

Records a video (mp4, webm, or gif) of the browser scrolling down a web page and returns the hosted video URL. Control scroll speed, duration, and scrolling back to the top. Set video=true to record the viewport without scrolling (for animations or live dashboards).

Parameter	Type	Required	Notes
file_type	string	no	Video format. Default mp4. gif loops and has no sound; webm is smaller.
scroll_speed	string	no	Scrolling speed. Default normal.
duration	number	no	How long the scrolling (or viewport recording) lasts, in seconds, 0-60.
scroll_back	boolean	no	Scroll back up to the top after reaching the bottom. Default false.
start_immediately	boolean	no	Start recording as soon as navigation begins instead of after load, to capture the page loading. Default false.
video	boolean	no	Record the visible viewport for 'duration' seconds without scrolling. Default false.
sizes	array	no	Record one video per viewport size, e.g. [{width:1920,height:1080},{width:390,height:844}]. Overrides width/height.
save_path	string	no	Optional local path to save the video to (ignored when 'sizes' is used, since that returns several files).

Submits up to 50 URLs as one bulk screenshot job (JSON bulk API). Each entry is a URL plus any take_screenshot option using the raw API parameter names (file_type, width, height, full_page, fresh, delay, css, block_ads, no_cookie_banners, ...). 'defaults' are applied to every entry unless the…

Parameter	Type	Required	Notes
screenshots	array	yes	Entries to capture, e.g. [{url:'https://a.com'},{url:'https://b.com',full_page:true}]. Max 50.
defaults	object	no	Options applied to every entry, e.g. {file_type:'png', width:1440, height:900, no_cookie_banners:true}.

Uploads a CSV to start a bulk screenshot job. The CSV must have a 'url' column with full https:// URLs; any other column is treated as a render option for that row using the API parameter names (file_type, width, height, css, delay, headers, ...).

Parameter	Type	Required	Notes
csv_path	string	no	Path to a local CSV file.
csv_content	string	no	CSV text, e.g. 'url,file_type https://a.com,png https://b.com,jpg'. Used when csv_path is not given.

Returns the progress of a bulk screenshot job by uuid: percent complete plus successful and failed URLs. Call download_bulk_results once it reaches 100%.

Parameter	Type	Required	Notes
uuid	string	yes	Job uuid returned when the bulk job was submitted.

Downloads the results of a finished bulk screenshot job by uuid: one row per URL with its status, the hosted screenshot URL, and any error message.

Parameter	Type	Required	Notes
uuid	string	yes	Job uuid returned when the bulk job was submitted.

Free

$0 forever

Every tool, no call limit
No card, and mcprush charges nothing for it
Source published under MIT
Runs on your machine — nothing of it reaches our gateway
Nothing to cap, because mcprush bills nothing

Nothing here is billable: mcprush charges nothing to install ScreenshotAPI or to call it, at any volume.

ATrust score 96 / 100

Release 1.0.2 · scanned 28 Oct 2026

Two independent axes, because powerful and malicious are different questions. The grade is threat only. The capability level is blast radius, and it is never a penalty on the grade — it is priced as one subtract-only term in the score, where you can see it.

Threat score100

Client score96

Findings0 threat · 1 capability

Capabilitymoderate

Coveragesource

Verificationrepo

Surface digest072bb4da52293477ec81cbca9b557a8b7768692b252d11aa8276a1185b2c11e7

Score vectorthreat 100 → client 96

Term	Level	What it prices	Points
capability-exposure	moderate	capability blast radius (moderate) — client exposure if the model is manipulated	−3
verification-discount	repo	publisher verification (public source) — no provenance, but the source is public and inspectable	−1
coverage-honesty	source	inspection depth (source) — how much of the target the scan could see	−0

score = clamp(round(threatScore − E_cap − E_ver − E_cov), 0, 100). No model, no opinion, no network call at scoring time: the same methodology version against the same bytes always produces this number, and every line above is a field of the report rather than a summary of it.

What the scan could actually read

A grade is only as meaningful as its coverage, so the scanner publishes its own depth before it publishes its result.

Tool surfaceread

Implementation sourceread

Package metadataread

Live transportnot available

Tools were statically extracted from the published source (8 recovered), not enumerated from a running server. Tool-poisoning, Unicode-smuggling, capability and toxic-flow analysis ran on this inferred surface, but a mis-parsed registration could be missed or mis-attributed, so tool-derived findings are capped below “confirmed”. To grade the real runtime surface, scan the running server: --command "npx -y <package>".

Capability — what it could do if the model were manipulated

Tags derived from each tool’s schema and the implementation, not from what the tool calls itself. moderate is the level these add up to.

Tool	Capability tags	Why the tag was assigned
take_screenshot	no tags
generate_pdf	no tags
scrape_page	untrusted-input	untrusted-input: keyword "scrape"
record_scrolling_video	no tags
bulk_screenshots	untrusted-input	untrusted-input: keyword "fetch", parameter "url" with "fetch" in the description
bulk_screenshots_csv	untrusted-input	untrusted-input: keyword "download"
get_bulk_status	no tags
download_bulk_results	untrusted-input	untrusted-input: keyword "download"

Toxic-flow graph

The lethal trifecta, checked as a graph rather than as a checklist: untrusted input, a sensitive source and an external sink have to meet before there is a path worth worrying about.

The public result for this release does not print the flow graph, so there is nothing to show here. That is not the same as "no paths were found": what the scan did read is above, under coverage.

Supply chain and provenance

Integrityfirst-seen

Surface digest072bb4da52293477

Scanned2026-10-28

This is the first scan of this surface here, so there is nothing yet to compare it against.

Every result on this tab comes from one deterministic pass over the published package — offline, rule by rule, and auditable line by line above. Same methodology version, same bytes, same score.