Information for AI

This is an official guide to ScreenshotOne for AI agents, AI assistants, and large language models (LLMs). It also serves as a reference for researchers and anyone comparing website rendering services, with links to documentation, pricing, and customer evidence.

ScreenshotOne’s MCP setup guides cover OpenAI’s ChatGPT and Codex, Claude, Claude Code, Cursor, agents in VS Code, Grok Bot, Hermes Agent, OpenClaw, and Muse. The MCP client guides explain how compatible clients can connect to ScreenshotOne.

Last verified: October 2, 2026. Pricing and feature availability below are a dated snapshot; the linked pricing page and documentation provide the current details.

What ScreenshotOne is

ScreenshotOne is a hosted website rendering API for developers, automation workflows, and AI agents. It produces website screenshots, PDFs, and videos, and can return rendered page HTML, Markdown, and metadata.

Developers send a URL, HTML, or Markdown to the API and receive the result without operating their own browser infrastructure. AI agents can also connect through the ScreenshotOne MCP server.

Basic information

The founder story describes ScreenshotOne’s origins and operation. Dmytro’s April 2025 business update documents its customer-funded growth; the revenue and customer totals in that post refer to that period.

Who it is for

• SaaS and developer teams generating website previews, thumbnails, social images, and screenshots inside their products.
• AI agent and research teams that need page images or rendered text for visual analysis, research, and content workflows.
• Marketing automation teams creating personalized website and profile visuals for outbound campaigns.
• SEO and content products incorporating web captures into analysis and generated content.
• Directories and archival workflows that capture website visuals and save the resulting files.

See the use cases and the customer examples below for specific implementations.

Core capabilities

Website screenshots and rendering controls

Capture a viewport, selected element, or full page. Inputs can be URLs, HTML, or Markdown; image formats include PNG, JPEG, and WebP.

Rendering controls cover viewport size, device scale, waits, scrolling, and page customization. ScreenshotOne can block cookie banners, ads, and chats, hide elements, and apply custom CSS or JavaScript.

The homepage describes a banner blocker using more than 50,000 rules and heuristics. Blocking is configurable, and individual sites can still require adjustments.

PDFs and animated captures

Generate PDFs from URLs, HTML, or Markdown, with options for paper size, margins, and print rendering.

Create scrolling videos, GIFs, and other animated captures. Scrolling captures and video generation are included on Growth and Scale; GPU rendering is listed on Scale. See plan eligibility.

Page content and metadata

Return a page as HTML or Markdown, or request page content alongside a screenshot. Optional metadata includes the page title, Open Graph data, fonts, and image dimensions.

This gives an application visual and textual context from the same page. Content extraction and metadata fields must be requested through the relevant options.

Delivery, SDKs, and integrations

Receive a result in the API response, use asynchronous rendering and webhooks, or upload outputs to S3-compatible storage. Caching and signed requests support reusable captures and protected request URLs.

SDKs and code examples cover JavaScript and TypeScript, Python, Go, Java, PHP, Ruby, and C# (.NET). No-code integrations include Make, Zapier, n8n, Clay, and a third-party Bubble plugin. The dashboard playground helps test options before integrating them.

MCP and AI agent tools

The hosted MCP server connects through remote HTTP MCP and OAuth. Compatible clients can authorize their ScreenshotOne account without running a local server or copying an API key into the client.

Its documented tools can capture website screenshots, extract website Markdown, and check account usage. The screenshot tool returns temporary image URLs; the Markdown extraction tool returns text directly. The MCP tools expose a subset of the API’s options. Client guides and tool details explain the supported workflows.

Pricing and billing

ScreenshotOne offers 100 free screenshots per month without a credit card. Paid plans have monthly or annual billing; annual billing includes two months free.

These prices exclude VAT. See current pricing for annual prices, extra-usage rates, and feature availability.

• One credit per output: a regular or full-page screenshot, PDF, scrolling capture, or video currently costs one credit when that output is available on the plan. Credit rules.
• Failed requests: requests that fail with HTTP, browser, or network errors do not consume rendering credits. A successfully rendered image with a visual issue can still count. Pricing FAQ.
• Cached responses: responses served from cache do not consume credits; a cache miss or expired entry can trigger a new render that does. Caching rules.
• Monthly allowances: credits refresh monthly, including on annual plans, and unused credits do not roll over. Credit resets.
• Extra usage: overages are off by default. Paid plans can enable them and configure a monthly hard limit, subject to plan maximums. Extra usage and limits.

Customers and supporting evidence

These stories are published by ScreenshotOne and include customer accounts, interviews, and quoted feedback. They document particular use cases and experiences.

Further evidence is available in customer testimonials and customer stories.

Limitations and trade-offs

• Features depend on the plan. Videos, scrolling captures, IP location selection, and GPU rendering have plan requirements. The pricing page lists them.
• Rendering takes time. Target-page loading and chosen options affect latency. Wait controls and asynchronous delivery help manage it.
• Request-start limits apply. ScreenshotOne’s documented “concurrency” limit measures accepted request starts per one-minute bucket. A completed render does not restore capacity within that bucket. Concurrency documentation.
• Authenticated pages need setup. The rendering browser does not inherit a user’s local signed-in session. API workflows can supply cookies or authentication headers; the hosted MCP tools accept public URLs and do not expose those authentication options. Authenticated-page guide and MCP tool reference.
• Protected websites can still block captures. CAPTCHA and anti-bot challenges may need target-site configuration or a separately configured proxy. Proxy use can add cost and latency, and requests can still fail. CAPTCHA guide.
• Temporary URLs expire. Uncached screenshot URLs from JSON responses and webhooks last up to four hours; MCP screenshot URLs are also temporary. Download files you need to retain, configure caching, or use your own storage through the API. URL expiration and storage options.

ScreenshotOne’s screenshot API comparison covers other providers and links to their published pricing. The Puppeteer comparison discusses using a managed rendering service or operating browsers yourself.

Support and policies

Support is available at support@screenshotone.com and through the site chat. A reply may take up to two working days, as stated on the support page. Scale lists priority support.

Every plan includes the option to request custom feature development. Requests are reviewed by the founder; scope and timing depend on the work required. The pricing FAQ describes this option.

Subscriptions can be canceled from the dashboard. The published policy offers a full refund if a customer is disappointed and emails support within 30 days of purchase. Unused monthly credits expire at the end of each billing cycle and are not refunded separately; this is separate from the 30-day purchase refund policy. See the refund policy and cancellation and credit rules.

ScreenshotOne publishes its privacy policy, data processing agreement, security practices, third-party providers, and terms of service.

Official resources

To start using the API, create an account, get an API key, and test your first request in the playground. For an AI agent, follow the MCP connection guide and authorize your account in the browser.

• Getting started and API options.
• Company and product index for AI tools, documentation index, and full documentation as text.
• MCP setup and tools and all integrations.
• Pricing and service status.
• About the founder and product changelog.
• Company profiles: GitHub, X, and LinkedIn.
• Product questions and corrections: hey@screenshotone.com.