# String docs


- **Get started**
- [Overview](/docs): Two APIs — get any page on the web, or build a feed from a sentence.
- [Quickstart](/docs/get-started/quickstart): Make your first request and get a page back in under a minute.
- [Authentication](/docs/get-started/authentication): Authenticate Web Access, Site integrations and Composer API requests with a Bearer API key.
- [Choosing an endpoint](/docs/get-started/choosing-an-endpoint): Fetch vs Search vs Browser vs Screenshots — pick the right tool.
- [Pricing](/docs/get-started/pricing): Flat monthly plans plus per-request usage, billed in USD.

- **Build with AI agents**
- MCP server
  - [Overview](/docs/mcp/overview): Web access for Claude, Cursor or any MCP client — search the web, fetch any URL with GET or POST and map sites, as clean Markdown or structured data.
  - [Remote (hosted) MCP](/docs/mcp/remote): Search the web, fetch any URL and map a site — the recommended hosted MCP server. Sign in, or bring an API key. Nothing to install.
  - [Self-hosted MCP](/docs/mcp/self-hosted): Search the web, fetch clean Markdown, map a site, or ask about String products from @usestring/mcp.
  - [MCP tools](/docs/mcp/tools): Search the web, fetch any URL with GET or POST and map sites as clean Markdown or structured data — parameters and return shapes for every Web Access MCP tool.
  - [MCP /v3 endpoint](/docs/mcp/v3): Connect to the /v3 MCP endpoint and see each tool's inputs, outputs, read or write classification, billing, errors and limits.

- **Web Access API**
- Fetch
  - [Overview](/docs/fetch/overview): Send a URL and get the page back as JSON, raw bytes or Markdown. The API picks rendering, proxy tier and captcha handling.
  - [Response formats](/docs/fetch/response-formats): Return a page as structured JSON, raw bytes, or clean Markdown.
  - [Structured extraction](/docs/fetch/structured-extraction): Pass a JSON Schema and get back only the fields you need, extracted with AI.
  - [JavaScript rendering](/docs/fetch/javascript-rendering): Pages that need JavaScript are rendered automatically. Set executeJS to always render a page in a browser.
  - [Waiting for page load](/docs/fetch/waiting-for-page-load): Wait for a rendered page to reach a load state, from its HTML being parsed to its network going idle, before it is read or before actions run.
  - [Iframes](/docs/fetch/iframes): Inline the content of a page's iframes, list them with their HTML, and run browser actions inside them.
  - [XHR capture](/docs/fetch/xhr-capture): Return the XHR and fetch() requests a rendered page makes, filter them down to the API calls you want, and wait for a specific response.
  - [Custom headers](/docs/fetch/custom-headers): Forward your own HTTP headers with a direct fetch.
  - [Proxies & geolocation](/docs/fetch/proxies-and-geolocation): Route a request through a specific country.
  - [Captcha solving](/docs/fetch/captcha-solving): Captcha solving is on by default. Disable it to fail fast instead.
  - [Screenshots](/docs/fetch/screenshots): Capture PNG or JPEG screenshots of a page, full-page or viewport.
  - [Browser actions](/docs/fetch/browser-actions): Drive a page with a sequence of clicks, typing, scrolling, and waits.
  - [Access control & KYC](/docs/fetch/access-control): How restricted destinations and request patterns are gated.
- [Search](/docs/search): Search Google, DuckDuckGo, Brave, or Mojeek and get the top organic results as structured JSON, or results pages as HTML.
- [Sitemap](/docs/sitemap): Discover every URL on a site from a single starting point.
- Site integrations
  - [Overview](/docs/sites/overview): Typed, versioned actions over a website's own API, such as Yelp business records, search and reviews.
  - [Making a request](/docs/sites/making-a-request): The integrations endpoints, how to read an action's schema, and a complete Yelp example.
  - [Responses & errors](/docs/sites/responses): The run envelope every action answers with, partial results, and what each status code means.
  - [Versioning](/docs/sites/versioning): Each site has its own semantic version, and the manifest and response envelope carry shape versions.
  - [Catalog](/docs/sites): Every site integration and the actions it offers.

  - **Sites**
  - [Yelp](/docs/sites/yelp): Yelp business records, search by term and location, and business reviews.
  - [Google News](/docs/sites/googlenews): Google News search with date, country and language filters.
  - [Amazon](/docs/sites/amazon): Amazon.com product search, category listings and product pages, and customer reviews by product, filtered by star rating and sorted top or most recent.
  - [Home Depot](/docs/sites/homedepot): Home Depot product reviews, every one a product page counts, ratings-only entries included, and products at one store with that store's price, stock, pickup and delivery availability.
  - [Google Flights](/docs/sites/googleflights): Google Flights search: itineraries with prices, airlines, flight numbers, legs and layovers.
  - [Google Hotels](/docs/sites/googlehotels): Google Hotels search by place and dates, and a property's details and every booking site's offers.
  - [Google Maps](/docs/sites/googlemaps): Google Maps place search by query and location, and a place's details and reviews by place id.
  - [Walmart](/docs/sites/walmart): Walmart.com search results priced and stocked for a chosen store or zip code, and product catalog records by item id.
  - [Ppomppu](/docs/sites/ppomppu): Ppomppu (ppomppu.co.kr) board listings, newest first, and posts with their full comment threads, votes and views.
  - [Axios](/docs/sites/axios): Axios (axios.com) stories with their full text, Axios Pro and membership stories included, and the newest stories across the site or within one vertical.
  - [The Wall Street Journal](/docs/sites/wsj): The Wall Street Journal (wsj.com) article metadata with the publicly shown opening paragraphs, each article's read-aloud recording, and the daily news archive.
  - [WeChat Official Accounts](/docs/sites/wechat): WeChat Official Account (公众号) articles: full text, account, author, publish time, summary, original or reprint, images and videos. Find articles by keyword, read them by link, or list an account's article collection.
  - [Bloomberg](/docs/sites/bloomberg): Bloomberg (bloomberg.com) news: full article text with headline, authors, times, tickers and section, the newest stories across the site, and keyword search.
  - [Caixin Global](/docs/sites/caixinglobal): Caixin Global (caixinglobal.com), the English edition of Caixin: the newest articles in a section, and articles with their full text, byline, section, time published and images.
  - [Chosun Ilbo](/docs/sites/chosun): Chosun Ilbo (chosun.com) news articles with their full text, a section's newest articles, and keyword search across Chosun's publications.
  - [Argaam](/docs/sites/argaam): Argaam (argaam.com) reports, daily market report data and articles from the Saudi financial portal.
  - [Reddit](/docs/sites/reddit): Reddit posts with their full comment trees, post search with subreddit, sort and time filters, and a subreddit's hot, new, top, rising or controversial listing.
  - [ChatGPT](/docs/sites/chatgpt): ChatGPT answers: the answer a signed-out visitor gets for a prompt with web search on, and the citations behind it.
  - [SHEIN](/docs/sites/shein): SHEIN (shein.com) product details (title, prices, images, colors, sizes with stock, category and rating) and keyword search.
  - [App Store](/docs/sites/appstore): Apple's App Store: an app's catalog record and its customer reviews by track id, on any one of Apple's storefronts.
  - [G2](/docs/sites/g2): G2 (g2.com) software product overviews and reviews, from the structured data G2 pages and its reviews partial publish.
  - [Google Play](/docs/sites/googleplay): Google Play app details (name, developer, category, rating, installs, descriptions, icon, content rating, price), and an app's reviews newest first.
  - [Instagram](/docs/sites/instagram): Instagram profile overview, post history, a single post, a profile's reels with their play counts, and a location's first feed page, as its logged-out web protocol answers.
  - [TikTok](/docs/sites/tiktok): TikTok user search, a profile's overview and video feed, one video's detail, a video's comments, and a hashtag's feed, as its logged-out web answers.
- Browser (CDP)
  - [Overview](/docs/browser/overview): Connect Puppeteer or Playwright to a remote Chrome over a CDP WebSocket.
  - [Browser contexts](/docs/browser/contexts): Save logins, cookies and site storage between browser sessions.
  - [Captcha solving in browser sessions](/docs/browser/captcha-solving): Have String solve captchas and challenge pages before your browser session's page loads them.
  - [Connect with Puppeteer](/docs/browser/puppeteer): Drive the remote browser with puppeteer-core.
  - [Connect with Playwright](/docs/browser/playwright): Drive the remote browser with playwright-core over CDP.

- **Composer API**
- [Overview](/docs/feed-builder/overview): Describe the feed you want in prose, answer Composer's questions, approve the result.
- [Quickstart](/docs/feed-builder/walkthrough): One real build from POST to published feed, with the actual requests and responses.
- [Builds](/docs/feed-builder/builds): Start, list, watch, steer and finish a feed build.
- [Gates](/docs/feed-builder/gates): How a build asks you a question, and how you answer it.
- [Streaming progress](/docs/feed-builder/streaming): Watch a build over server-sent events, and what resuming can and cannot recover.
- [Feed data](/docs/feed-builder/feeds): List the feeds your organization subscribes to and download the files each one has collected.
- [Errors](/docs/feed-builder/errors): One envelope, the status mapping, and the two refusals that are not what they look like.
- [Limits](/docs/feed-builder/limits): Check the Composer API's limits before you design an integration: delivery, resuming a stream, approvals, billing, key revocation and request sizes.

- **Guides**
- [Markdown for LLMs](/docs/guides/markdown-for-llms): Turn web pages into clean, token-efficient Markdown for AI pipelines.
- [Structured data extraction](/docs/guides/structured-data-extraction): Get clean, typed fields out of a page with a JSON Schema.
- [Browser automation](/docs/guides/browser-automation): Drive interactive pages with actions or a full CDP session.
- [LangChain](/docs/guides/langchain): Give a LangChain agent web search and clean Markdown page fetches through the Web Access API.
- [AI SDK](/docs/guides/ai-sdk): Give a Vercel AI SDK agent web search, clean Markdown page fetches and site maps through the Web Access API.

- **Integrations**
- [Overview](/docs/integrations/overview): Use String from the automation platforms you already run, with no SDK and no custom code.
- [n8n](/docs/integrations/n8n): Call String from n8n's HTTP Request node — no community node to install.
- [Zapier](/docs/integrations/zapier): Add the String Web Access app to a Zap, or call String from a Webhooks by Zapier action.

- **Resources**
- [Support](/docs/resources/support): Get help and request access.
- API Reference
  - [Overview](/docs/api-reference/overview): Base URL, authentication, content type, and response headers.
  - [POST /fetch](/docs/api-reference/fetch): Full request and response reference for the Web Access API fetch endpoint.
  - [POST /search](/docs/api-reference/search): Search Google, DuckDuckGo, Brave, or Mojeek and return the top organic results, or results pages as HTML.
  - [Sitemap jobs](/docs/api-reference/sitemap): Full request and response reference for the asynchronous sitemap crawl endpoints.
  - [GET /wss](/docs/api-reference/browser-websocket): Connect to a browser over CDP, optionally restoring a saved profile.
  - [Errors](/docs/api-reference/errors): Status codes and error response shapes.
  - [Rate limits and concurrency](/docs/api-reference/rate-limits): Request rate limits, concurrent fetch and browser limits, and how to handle 429s.
  - [API versioning and deprecation](/docs/api-reference/versioning): How the Web Access API changes, what counts as breaking, and how an endpoint or field is deprecated before it is removed.