About WebReaderAPI

What problem does WebReaderAPI solve?

Language models read text, but web pages are mostly something else: navigation, scripts, cookie notices, ads, related-post widgets and footers. Feeding raw HTML to a model wastes its context window and buries the facts.

WebReaderAPI fetches the page, finds the main content and returns it as Markdown, with headings, lists, tables and links kept in place. You also get the title, author, publication date, language and word count when the page states them. One call replaces a fetch step, a browser step and a cleaning step in your own code.

How does a request work, step by step?

1. Your code sends a URL to https://webreaderapi.com/api/v1/read with your API key.
2. We check the site's robots.txt for our token and for all robots, and look for an ai-input=no signal.
3. If the page is allowed, our bot fetches it with the user agent WebReaderAPI/1.0 (+https://webreaderapi.com/bot/).
4. When the raw HTML holds too little text, a headless browser renders the page first (render=auto).
5. We extract the main content, convert it to Markdown and send it back with its metadata.

1 credit = 1 page read without a browser; 2 credits when the page needs a browser to render JavaScript. A request that fails, or that a site's robots.txt refuses, is not charged. The read endpoint documents every parameter.

What will WebReaderAPI never do?

Some limits are part of the product, not settings:

• It never reads a page a site has refused. We respect robots.txt: our crawler identifies itself as WebReaderAPI, follows the rules written for it or for all robots, and does not read a page that a site disallows. Refused pages are not charged.
• It never works around a refusal. If a site refuses our request, we return its answer: we never solve CAPTCHAs, log in to accounts, or try to get around a block or a paywall.
• It never reads behind a login or a paywall, even with credentials you provide.
• It never keeps your pages. The text goes back to you, and our logs hold only the host name.

These rules cost us some pages we could technically fetch. We accept that trade, because site owners, customers and AI assistants all need to trust the tool.

Who are the customers?

Mostly developers who give AI agents a reliable way to read the web, and teams that load documentation, help centres or blogs into retrieval systems. Researchers use it to collect public articles in a clean form, and some tools use it to build reading views or summaries.

The Acceptable Use Policy lists what is not allowed, such as collecting personal data about private people. A person looks at every report, and accounts that break the rules are suspended.

Why crypto and no KYC?

Because many of our customers are individual developers, small teams and autonomous projects that prefer not to share identity documents or card details with every API they try. Account-number login; email is optional. You top up from $25 in BTC, USDT, USDC, ETH, SOL, LTC, then buy credits from your balance.

No KYC does not mean no rules. The acceptable use policy applies to every account, robots.txt applies to every request, and we act on documented abuse reports.

What do we not claim?

We publish no response-time, uptime or success-rate figures today. Success depends on the sites you read, and an honest figure needs a measured public sample with a method and a date. When we have one, it will appear on the status page.

We also do not claim to read every page on the web. Sites that refuse robots, require a login or block automated requests stay out of reach by design.

How can I reach a person?

Write to [email protected] for questions about the API, billing or your account; a person answers, every day of the week. Site owners and anyone reporting misuse can write to [email protected]; the bot page explains how to block or slow our crawler.

Nobody from our team will ever ask you for your account number or an API key.