Hand your AI a URL, get the page back as clean Markdown.
Give it the URL of a public web page and it returns the page’s content as Markdown or plain text, ready for a model to read. Use it from Claude, ChatGPT, Cursor, Codex or your own code — no scraping service account of your own, quoted before each call, paid from one balance.
1Connect your AI app
Pick the app you use. You do this once, and every API on toolfare is then available in it.
Add toolfare as a connector in Claude
Add to Claudehttps://api.toolfare.ai/mcp
- 1Open Customize → Connectors and choose "Add custom connector".
- 2Paste this URL as the remote MCP server URL.
- 3Leave the OAuth client id and secret under Advanced settings empty.
- 4Complete the browser authorization, then return to your conversation.
Custom connectors are available on Free, Pro, Max, Team and Enterprise. A Free account may hold one custom connector in total.
Claude’s own instructions2What to ask
Copy a prompt into your AI app once you’re connected. It picks the right one of the 1 tools and tells you the price first.
One page, as Markdown or plain text
From $0.021Read a public web page as Markdown or plain text
- “Read this blog post and summarise it in five bullets.”
- “Turn this documentation page into Markdown I can paste into our wiki.”
- “Read the pricing page at this URL and list each plan’s limits.”
1Create an API key
One key and one endpoint for every API on toolfare.
export TOOLFARE_API_KEY='tf_live_…'
2Every tool and its input
Send the tool name and its arguments as JSON; the reply is JSON too.
web_reader_readRead a public web page as Markdown or plain text
From $0.021
Required parameters:
urlOptional parameters:
format
Use it when
- You have a URL and want the agent to work from the page’s actual text.
- You need a page in Markdown for notes, a prompt or a retrieval index.
- Your agent found pages with a search and now needs to read the best one.
Look elsewhere when
- You need many pages of the same site, not one. Use this instead
- You do not know the URL yet. Use this instead
- You need to see how the page looks rather than read its text. Use this instead
Web Reader directly, or through toolfare
Directly
- A scraping or reader API account and key, or your own headless browser
- Code to strip navigation and boilerplate and convert HTML to Markdown
- Handling for timeouts, redirects and pages that block plain HTTP clients
Through toolfare
- One call per page, quoted before it runs
- No scraper to host and no vendor key in an agent’s config
- Search, crawl and answer tools behind the same connection
toolfare passes the page through as the reader returns it; it does not reach private pages or clean up a page any further.
Limits
- One URL per call.
- Public pages only — content behind a login or paywall cannot be read.
- Text output: Markdown (the default) or plain text. Images and layout are not returned as visuals.
- Pay as you go from your toolfare balance, in US dollars; there is no subscription. A call runs when your balance covers its quote.
Questions
A call that takes a web page’s URL and returns its main content as Markdown, which language models read more reliably than raw HTML. Here it is `web_reader_read`.
Markdown keeps headings, lists and links, which helps a model follow the structure. Plain text is shorter when you only need the words.
No. It reads public pages only, and it never uses your credentials for any site.
Each tool’s current price is shown on this page and quoted before the call runs.
Yes. Call it over HTTP with a toolfare API key, or from a script with the toolfare CLI.