Give your agent
a clearer view.
Turn public webpages into useful text—with a source, a timestamp, and an explicit limit. A focused reading tool for agent workflows.
0.01 USDC per successful extraction.
Send a URL
One endpoint. Choose how much text your agent needs.
Get usable text
Keep page details while removing common navigation and recognized notices.
Keep the source
Every result includes its origin, retrieval time, and truncation status.
Made for a simple tool call.
POST /v1/read
Content-Type: application/json
{
"url": "https://example.com/article",
"max_chars": 8000
}A payment-enabled client receives an x402 price quote, authorizes payment within its budget, and retries the request. Normal API clients can inspect the quote without paying. The client must support the advertised network and asset.
Clear capabilities. Clear limits.
| Input | Public HTTP/HTTPS HTML and plain text |
| Output | Readable text, title, source URL, timestamp and truncation flag |
| Output size | 8,000 characters by default; up to 50,000 |
| Page size | Up to 2 MB decoded content |
| Not supported | JavaScript rendering, PDFs, logins or CAPTCHA bypass |
| Extraction | Heuristic; some complex pages may return partial content |
A small tool with an honest contract.
No browser or AI model runs on the static extraction path. The service prepares readable content before requesting payment settlement. Failed extraction is not submitted for settlement. Settlement timeouts can have an uncertain outcome; retries reuse the same authorization rather than requesting a new payment.
Webpage content is untrusted source material, not instructions for your agent. Use sources you have permission to access. Payment records retain request fingerprints and results temporarily to support safe retries; do not submit sensitive URLs.