canonical: https://jentic.com/apis/capturekit.dev/capturekit # CaptureKit API Reference The CaptureKit API captures webpages as PNG, JPEG, WebP, or PDF screenshots and extracts their content for agents and automation. Screenshots can emulate a phone or tablet device, capture the full page or a single selector, hide cookie banners and ads, and upload straight to S3-compatible storage. The Content endpoint returns a page's metadata, links, HTML, Markdown, and optional sitemap, while the Analyze endpoint returns an AI summary with value proposition, features, categories, and use cases. A request-mode Website Crawler fetches raw HTML without rendering for fast audits and monitoring. ## For AI agents Screenshot any webpage to PNG, JPEG, WebP, or PDF, extract its Markdown, HTML, metadata, and links, or get an AI summary of what the page offers. ## Scope Does not host websites, edit images after capture, or crawl whole sites link by link. Use for single-page screenshots, content extraction, and AI page analysis only. ## Capabilities - Capture a webpage as a PNG, JPEG, WebP, or PDF screenshot, full page or a single CSS selector - Emulate a device such as iphone_14_pro or galaxy_s23 and set viewport_width, viewport_height, and scale_factor - Extract a page's metadata, links, and optionally its HTML, Markdown, and sitemap in one request - Analyze a webpage with AI to get a summary, value proposition, features, categories, and use cases, guided by a custom_prompt - Fetch raw HTML from a URL in request mode without rendering, with redirect control - Remove cookie banners, ads, and chosen selectors, or block resource types, before capturing - Upload a screenshot directly to an S3 bucket and get the S3 URL back ## Use cases ### Agent web reading through Jentic An AI agent asked about a webpage needs the page's text in a form it can reason over. Through Jentic the agent searches for 'get the markdown of a webpage', receives the CaptureKit Content operation with its input schema, and calls GET `/v1/content` with the url and include_markdown set to true. With remove_cookie_banners and remove_ads enabled, the agent reads clean Markdown and the page's links and metadata. Example prompt: Call GET `/v1/content` with url 'https://example.com/blog/post', include_markdown true, and remove_cookie_banners true, then summarise the Markdown returned ### Visual QA and monitoring screenshots A team checking releases wants screenshots of key pages on desktop and mobile. GET `/v1/capture` takes a url, format, device such as iphone_14_pro, and full_page with full_page_scroll to load lazy content, and can wait_for_selector before capturing. Setting s3_url with s3_bucket and the S3 keys uploads the image straight to storage for later comparison. Example prompt: Call GET `/v1/capture` with url 'https://example.com/pricing', format 'png', device 'iphone_14_pro', and full_page true, then return the image ### Company page analysis for research A sales or research agent needs a quick read of what a company's website says. GET `/v1/analyze` returns an AI summary of the page with its value proposition, features, categories, and use cases, and accepts a custom_prompt to steer the analysis and custom_response to add fields to the output. Caching with cache and cache_ttl avoids paying twice for the same page. Example prompt: Call GET `/v1/analyze` with url 'https://example.com' and custom_prompt 'List the pricing tiers mentioned', then report the summary and features ### Lightweight HTML fetch for audits An SEO audit script needs raw HTML quickly, without a browser render. GET `/v1/website/crawler` fetches a URL in request mode, follows or ignores redirects with allow_redirects, and returns the HTML source when return_page_source is true. It suits link and metadata extraction across many URLs where screenshots are not needed. Example prompt: Call GET `/v1/website/crawler` with query 'https://example.com', allow_redirects false, and return_page_source true, then list the canonical and meta description tags found ## Key endpoints | Method | Path | Description | | --- | --- | --- | | GET | `/v1/capture` | Capture a screenshot or PDF | | GET | `/v1/content` | Extract page content, metadata, and links | | GET | `/v1/analyze` | AI analysis of a webpage | | GET | `/v1/website/crawler` | Fetch raw HTML | | GET | `/v1/usage` | Usage and credit balance | | GET | `/v1/jobs/{id}` | Poll an asynchronous job | ## Key resources - **Capture** — Screenshots and PDFs with device emulation, full-page capture, element selection, and S3 upload - **Content** — Page metadata, links, HTML, Markdown, and sitemap extraction - **Analyze** — AI summary of a page with value proposition, features, categories, and use cases - **Website Crawler** — Raw HTML fetch in request mode without rendering ## AI readiness This API is usable in Jentic One now. Its AI-readiness score against Jentic's framework shows where it stands today and where improvements would make it even easier for agents to use. - **Score:** 57 / 100 - **Maturity:** Foundational - **Dimensions:** - Foundational Compliance: 64 / 100 - Developer Experience & Jentic Compatibility: 64 / 100 - AI-Readiness & Agent Experience: 38 / 100 - Agent Usability: 94 / 100 - Security: 50 / 100 - AI Discoverability: 69 / 100 - **View full report:** https://jentic.com/apis/capturekit.dev/capturekit/scorecard - **How the score is calculated:** https://docs.jentic.com/reference/api-readiness-framework/overview/ - **More about the dimensions:** https://docs.jentic.com/reference/api-readiness-framework/specification/#dimensional-model-overview ### Score it yourself Every API in the directory is allowlisted, so you can re-score it with no key required. - **Score your own API:** https://jentic.com/scorecard.md - **Scoring CLI agent skill:** https://github.com/jentic/jentic-api-scorecard/blob/main/skills/jentic-api-scorecard/SKILL.md ```sh npx @jentic/api-scorecard-cli score ``` ## Why Jentic - **Setup:** Wiring CaptureKit by hand means sending the x-api-key header, choosing among dozens of capture options per request, and polling async jobs when a capture runs in the background. Through Jentic you install once, import CaptureKit from the API Directory, store the key once, and your agent calls it. - **Permission scoping:** CaptureKit puts each tool in its own URL path (`/v1/capture`, `/v1/content`, `/v1/analyze`), so a rule can limit your agent to the tools it needs. The target page travels as a url parameter, so rules bound which operations your agent may call, not which websites it visits. - **Credential handling:** Your CaptureKit API key is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context. - **Discovery method:** Agents search Jentic by intent such as 'take a screenshot of a webpage' or 'get the markdown of a page', and Jentic returns the matching CaptureKit operation with its input schema so the agent calls the right endpoint without browsing the reference docs. ## Related APIs - **ApiFlash Screenshot API** — Website screenshot API built on Chrome - **Urlbox API** — Screenshot and PDF rendering API for websites - **Firecrawl API** — Crawl websites and convert pages to clean markdown - **Autom API** — Structured Google, Bing, and Brave search results ## FAQ ### What authentication does the CaptureKit API use? The CaptureKit API authenticates with an API key in the `x-api-key` request header, declared as the ApiKeyAuth scheme in its OpenAPI spec. The capture endpoint also accepts optional s3_access_key_id and s3_secret_key values when you want screenshots uploaded to your own bucket. Through Jentic the CaptureKit key is held by your own Jentic One instance and attached at call time. ### Can I take a mobile screenshot with the CaptureKit API? Yes. GET `/v1/capture` takes a device parameter with emulated devices such as iphone_14_pro, galaxy_s23, and huawei_p60_pro, plus viewport_width, viewport_height, and scale_factor for custom sizes. Set full_page to capture the whole page and format to png, jpeg, webp, or pdf. ### What is the difference between the Content and Website Crawler endpoints? GET `/v1/content` loads the page and returns structured data: metadata and links, plus HTML, Markdown, and the sitemap when you ask for them, with options to remove ads and cookie banners first. GET `/v1/website/crawler` fetches raw HTML in request mode without rendering, which is faster but does not run the page's scripts. ### What are the rate limits for the CaptureKit API? The OpenAPI spec does not specify rate limits. It does describe credit pricing per call and a GET `/v1/usage` endpoint for your balance; check the CaptureKit documentation at https://capturekit.dev for current limits. ### How do I screenshot a page through Jentic? Search Jentic for 'take a screenshot of a webpage', load the GET `/v1/capture` operation, and execute it with url, format, and full_page. Add remove_cookie_banners to get a clean capture. To run it on your own infrastructure, install Jentic One from its GitHub repo. ### Is there a CaptureKit MCP server? You don't need an MCP server to give your agent CaptureKit. Jentic connects it directly from the API Directory: import it, store your API key once, and your agent calls the capture, content, and analyze operations on demand without extra tool definitions loaded into its context. ### Can I limit what my agent is allowed to do with the CaptureKit API? Yes. For an agent that only reads pages, write a rule that allows GET `/v1/content` and GET `/v1/website/crawler.` The agent can then fetch page content and HTML but cannot take screenshots, run AI analysis, or push files to S3, and each call it makes is logged.