Service eligibility and regional restrictions
PuppyIP serves only compliant overseas businesses and their authorized personnel. Proxy services are not available in mainland China. The service may only be used for lawful business activities outside mainland China. Use of this service within mainland China is prohibited.
Hosting a proxy IP or server overseas does not change these restrictions. The service must not be provided to end users in mainland China through relaying, forwarding, sharing or resale. Before use, read the Terms of Service.
Key Takeaways
- OpenAI created the incident at 22:50:17 UTC on November 11, 2026 and said it was investigating a problem starting new runs.
- At the 23:41 UTC check, the status remained investigating and the latest announcement was still timestamped 22:50:17 UTC.
- The notice did not identify affected accounts, plans, countries, models or existing runs, or provide a cause, workaround or recovery estimate.
- Check the same official incident and your original task’s outcome. Loading a page, clicking a button and actually starting a run are separate results.
Which startup problem has OpenAI confirmed?
OpenAI’s official status record confirms an issue preventing users from starting new Workspace Agent runs. A run is an attempt to execute a task. The announcement concerns new runs and does not establish whether tasks already underway are affected. If an error looks similar, first check the product and the step that failed; the word “Agent” alone does not connect another product’s error to this incident.
The incident is marked “Investigating.” This acknowledges a platform issue with starting runs, but it does not establish the cause of every failure you see or show that every account is unable to use the service.
When was the announcement made, and when was it checked?
The incident was created at 22:50:17 UTC on November 11, 2026, or 06:50:17 on November 12 in Beijing. Its latest update carries the same timestamp. OpenAI said it was investigating and would provide further information when available.
When this article checked the official incident list at 23:41 UTC on November 11, the record still showed investigating. That was 07:41 on November 12 in Beijing; this article’s publication date uses Beijing time. The check time is separate from the actual onset of the problem, and the announcement’s creation time does not establish when an individual user was first affected.
What information should you preserve now?
As a troubleshooting suggestion, note the task you intended to start, the attempt’s time and time zone, the exact error, and whether a run record or output already exists. Remove tokens, private task contents and account identifiers before sharing diagnostic material. This lets you compare the same attempt later, rather than substituting a successful home-page visit on another device.
If an earlier attempt could have produced output or performed an external action, inspect its original record before deciding whether to start over or retry. Clicking Start does not establish that a run began, and a stuck page does not prove that nothing executed in the background. This is advice to avoid duplicate actions, not a claim that OpenAI has confirmed duplicate execution in this incident.
Which causes and recovery steps remain unknown?
The announcement does not give a root cause, fix or estimated recovery time. It also does not list affected plans, regions, models or account counts. The critical impact label does not supply those missing details; a region’s absence from the notice does not establish that it is unaffected.
The update does not attribute the issue to a user’s network, IP address or usage, or say that a client upgrade, workspace rebuild or different network exit will fix it. Avoid changing an existing workspace, permissions or business connection simply to test an unconfirmed remedy. Preserve evidence of any other errors so they can be investigated separately.
How can you tell when work can continue?
For later official updates, keep checking the same incident about starting new Workspace Agent runs, alongside your original task. Record any new status and its announcement time. The current investigation notice does not say that a cause has been identified, a fix applied or service restored; those conclusions require further evidence.
If OpenAI later confirms recovery, use your team’s existing process to check whether the original attempt completed and which work remains unfinished. A new run starting, the task completing and its external results being correct are separate checks. The official status and your task’s own record each provide information; neither replaces the other.
Sources
Frequently Asked Questions
Should I delete and recreate the workspace?
The current official notice does not prescribe that recovery step. Preserve the existing workspace and original task records first. If OpenAI provides separate guidance for your specific error, check its scope before applying it.
Does this announcement include refunds, compensation or extra usage?
No. The update only acknowledges an investigation into new-run startup failures. It does not announce fee changes, compensation, refunds or additional usage; those require separate, explicit official information.