Bot signals · AI crawler & agent
What is Cursor Cloud Agents?
Cursor Cloud Agents are Cursor’s coding agents that run in isolated virtual machines in the cloud. These requests come from them when they clone and push to a customer’s repositories, download packages, call external APIs such as GitHub’s or open web resources while they work.
01 · Profile
Cursor Cloud Agents at a glance.
What to look for in your logs, and what the lookup API returns for its addresses.
- Type
- AI crawler & agent
- API flag
risk.is_ai
A user agent alone never sets a flag. Anyone can claim to be Cursor Cloud Agents; the flag follows the address the request comes from.
02 · In the response
What a match looks like.
A lookup of a Cursor Cloud Agents address sets risk.is_ai. The flags come with geolocation and network data in the same response, for API keys on every paid plan.
The response says which kind of bot an address belongs to, not which operator. A request that claims to be Cursor Cloud Agents while risk.is_ai is false should not be trusted as Cursor Cloud Agents.
{
// … location and network fields …
"risk": {
"is_ai": true,
"is_crawler": false,
"is_scanner": false
}
}
03 · Control
Allow or block Cursor Cloud Agents.
Your options
- robots.txt works only for crawlers that choose to follow it. Check the operator’s documentation for the token it reads.
- Firewall, WAF or CDN. Many offer bot rules you can allow or deny by category. Use them to enforce what robots.txt only asks for.
- In your app. Look up the client address and allow or deny on
risk.is_ai, alongside the user agent. - Never on the user agent alone. Blocking the name stops only honest clients; allowing it lets anyone in who borrows it.
04 · Questions
About Cursor Cloud Agents.
What is Cursor Cloud Agents?
Cursor Cloud Agents are Cursor’s coding agents that run in isolated virtual machines in the cloud. These requests come from them when they clone and push to a customer’s repositories, download packages, call external APIs such as GitHub’s or open web resources while they work.
How do I block Cursor Cloud Agents?
Deny it at your firewall, WAF or CDN, or look up each client address and deny requests where risk.is_ai is true. Do not rely on the user agent alone.
Can I trust a request that says it is Cursor Cloud Agents?
Not on the user agent alone: any client can send a bot’s name in its user agent. A user agent alone never sets a flag. Check the address with the free bot checker or the lookup API.
Which risk flag does Cursor Cloud Agents set?
Its addresses set risk.is_ai in the lookup response. The risk flags come with every paid plan.
Related bots
Recognise Cursor Cloud Agents on every request.
is_ai, is_crawler and is_scanner in every lookup, on every paid plan, from 9,00 € a month (VAT included).