AI incidents: Cursor and AI Studio recoveries, ChatGPT loading and runtime reports — 2 November 2026

Publisher disclosure: MUVO offers competing music, remix, and video tools. This report summarizes public sources and provider documentation; it is not an independent product test or a claim that MUVO is free of similar issues. Read our reporting method.

Incident summary and status at cutoff

Cursor has resolved its billing and usage dashboard incident. Google lists a resolved AI Studio API-key outage, while fresh customer and support updates describe recurring ChatGPT chat-loading and Python/container failures. The dashboard events are provider-confirmed; the ChatGPT cases are account-level reports with relevant support corroboration.

Coverage runs from 30 October at 14:21 through 2 November at 11:59 Europe/Istanbul (UTC+3), with earlier unresolved cases rechecked. This edition also includes findings held from the unpublished 1 November draft. Their original dates are retained; they are not presented as new outages today. Releases and forthcoming changes are in the separate news edition.

Cursor: billing and usage dashboard incident resolved

The official SpaceXSI status page marks Cursor’s billing and usage dashboard degradation resolved at 13:53 Istanbul on 1 November. It names cursor.com and Enterprise Dashboard as affected components. Investigation began at 08:38, followed by a continuing-investigation update at 11:44. Official incident and timeline.

The notice was open for approximately 5 hours 15 minutes, measured from its first investigation post at 05:38:43 UTC to the resolution post at 10:53:33 UTC. This is the acknowledgement-to-resolution interval, not a verified measure of the actual impact duration: the underlying impact onset is unknown.

The provider gives no cause, region, workaround or detailed mitigation. Dashboard degradation does not establish incorrect charges, failed payments or a Grok model/API outage. Provider resolution cannot confirm every account’s recovery. This is separate from the resolved workflow disruptions in the 31 October edition.

Google AI Studio: API-key traffic outage marked resolved on 30 October

Google’s AI Studio history acknowledges an outage affecting all traffic using API keys in AI Studio, then marks it resolved. The detection notice says requests were failing and investigation was underway; the resolution supplies no additional explanation. Original AI Studio and Gemini API status history.

Both entries are dated 30 October. The page displays 14:42 for detection and 17:18 for resolution without specifying a timezone. We preserve those source clock readings and do not convert them to Istanbul time or calculate an outage duration. Exact impact onset, cause, mitigation and region remain unknown.

The confirmed scope is API-key traffic in AI Studio. It does not establish an outage of the consumer Gemini app, every Google AI service or every authentication method. This recovery was verified for the pending 1 November draft and rechecked today.

ChatGPT: saved-chat loading errors recur after temporary improvement

OpenAI Support acknowledged saved-chat “Too many requests” failures at 19:11 Istanbul on 1 November, asking the customer to check for HTTP 429. Support response.

At 01:56 Istanbul on 2 November—22:56 UTC on 1 November—the customer reported recurrence after several working days and HTTP 429 in Chrome after switching windows. That is not evidence of uninterrupted downtime since their original 28 October report. Customer update.

A separate customer posted at 21:47 Istanbul on 31 October that saved chats failed in Chrome and Firefox on Windows while remaining accessible in the iPhone app. The customer reports that conversation listings succeeded but retrieving contents returned HTTP 429. Independent account.

These updates continue the issue discussed in our 27 October incident report. They do not establish a common cause across accounts, widespread data loss, a universal iPhone workaround or a company-wide outage. General recovery and affected-account counts remain unverified.

ChatGPT: Python and container failures return in a customer workflow

At 13:49 Istanbul on 1 November, a customer reported renewed caas.internal.errors.ClientError failures after successful work and an overnight break. Two runtime checks failed in the same conversation. This updates that author’s 30 October report; those posts are not independent accounts. Recurrence update.

Independent earlier accounts and 26 October OpenAI Support replies corroborate related problems. Support described a mitigation for some attachment timeouts, including more time for files to become available on a specific account—not a universal fix. Its applicability to this recurrence is unknown. Support replies, independent account.

Status: customer-reported recurrence; general recovery, cause and continuous duration unverified. File visibility alone does not demonstrate usable Python/container execution.

What affected users can check

For saved-chat loading failures, the provider’s public support reply asks for the error text or HTTP status code. Keep conversation links, logs and account details within the provider’s private support process. A working mobile session in one customer’s account is useful counterevidence to total account loss, not a guaranteed recovery step for everyone.

For Python/file-processing failures, support requests the retry time and timezone, exact error and available request identifier through private support. A successful attempt alone does not demonstrate lasting recovery.

Sources

• Cursor: official billing/usage incident.
• Google: AI Studio status history.
• OpenAI community: saved-chat support acknowledgement, recurrence, independent web/iPhone comparison, runtime recurrence and earlier mitigation response.

Official histories and feeds for Anthropic, OpenAI, Google and ElevenLabs were checked separately, alongside the wider creative-tool watchlist. No newly verified Anthropic or ElevenLabs event was added. That is a finding about the accessible evidence, not proof that every service or account worked. Community posts are not verified-purchase reviews, and small samples cannot establish prevalence. Provider-resolution claims and customer recovery remain distinct.

MUVO publishes this reporting and offers competing creative tools. See our reporting method.

Update timeline

• 2 November 2026, 11:59 Istanbul / 08:59 UTC: Research cutoff. Added Cursor’s 1 November resolution, retained the dated AI Studio recovery, and incorporated fresh ChatGPT support and recurrence evidence. Pending findings from the unpublished 1 November draft are included with their original dates.