Terms of Use
Last updated July 19, 2026.
AgentRead is an early-stage product. By using it you agree to the following, written in plain language on purpose.
What the service does
AgentRead fetches a URL you provide, extracts its readable content, and returns clean Markdown plus a ReadScore. You're responsible for only submitting URLs you have the right to fetch and process — don't use this to bypass a paywall, access-control system, or robots.txt disallow rule you don't have permission to bypass.
No warranty
The ReadScore and flags are a heuristic, not a guarantee — they're meant to help you find real issues, not to certify that a page is or isn't safe for an AI agent to rely on. The service is provided "as is," without warranty of any kind. We are not liable for decisions made based on its output.
Fair use / rate limits
The anonymous Playground and ReadScan tools are rate-limited per IP; the authenticated Read API and MCP server are rate-limited per API key. We reserve the right to throttle, suspend, or revoke access that abuses these limits or that we reasonably believe is being used to scrape content at scale in a way that harms target sites.
Accounts
You're responsible for keeping your API keys confidential. If a key is compromised, revoke it from your dashboard immediately — we do not monitor for key misuse in real time today.
Changes
This is a young, fast-moving product — these terms, and the product itself, may change. We'll update the date at the top of this page when they do.
Contact
Questions: reach out via the contact details on our GitHub repository.