API Error - "We encountered an unexpected error repeatedly" when calling Cloud Agent

Support Bug Reports
,
1

Dear Cursor Support Team,

I am encountering a persistent error when trying to use the Cursor Cloud Agent API through the SDK, and I would appreciate your assistance in diagnosing the issue.

Issue Summary:
Whenever I attempt to make an API call, the agent returns the following error:

“We encountered an unexpected error repeatedly. You can retry by writing a follow-up below.”

The error occurs consistently and I have not been able to successfully complete any API request.

My Setup:

  • Subscription Plan: Pro ($20/mo) – active and paid

  • API Key: Created and valid (crsr_...)

  • SDK: Cursor TypeScript SDK

  • Usage Context: Calling from a third-party application / local script

  • Agent Type: Cloud Agent (via SDK configuration)

  • OS: Windows

Attached Information:

  • Request ID / bcId: bc-8ec8b9b1-a5d6-403d-9a12-7fb7e1076d96

  • Deep Link: cursor://anysphere.cursor-deeplink/background-agent?bcId=bc-8ec8b9b1-a5d6-403d-9a12-7fb7e1076d96

  • Time of error: July 15, 2026

  • Cursor version: Windows (latest)

  • Subscription: Pro – auto-renews on July 27, 2026

What I’ve Already Tried:

  • Retried multiple times after waiting 5–10 minutes

  • Verified my API Key is correct and has Admin scope

  • Checked the Dashboard – API usage still shows 0.0%, confirming no successful calls have gone through

I believe this is related to the known Cloud Agent service-side issue mentioned in the community forum. However, I would like to confirm whether this is the same root cause or if there is something specific to my account/configuration.

Please let me know if you need any additional logs or information to help troubleshoot this.

Thank you for your time and support. I look forward to your reply.

Best regards

5

Hey, thanks for the detailed report with the bcId. That really helped.

For this request bc-8ec8b9b1-a5d6-403d-9a12-7fb7e1076d96, we can see the failure was on our side. It was a transient incident in the Cloud Agent infrastructure around the time you started the run. It’s not related to your API key or account. Your key and scope look fine.

Good news: the incident has settled down, and we haven’t seen similar errors in the last few hours. Please try again, it should go through.

A couple quick notes:

  • 0% usage on the dashboard is expected. The run failed at startup and didn’t execute anything, so no tokens were used.
  • If the error happens again after retrying, send a new bcId and I’ll check that one separately.

Let me know how it goes.