AI incidents: OpenAI, ElevenCreative, Runway and Grok recoveries — 31 October 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

OpenAI, ElevenLabs and Runway report recovery from search, image-edit and generation failures. The official Grok and Cursor history adds seven separate resolved disruptions. These eleven events were acknowledged by their operators; none is inferred solely from a complaint graph or a current status indicator.

This edition covers 28 October at 12:06 through 31 October at 11:55 Europe/Istanbul (UTC+3). Some Grok and Cursor notices were newly accessible during this scan but describe earlier events inside that window. They are not new 31 October outages.

These are notice timestamps, not established boundaries of user impact. Exact onset and full affected populations are not supplied. We do not turn the gaps between notices into measured outage durations. A provider’s resolved status does not establish recovery for every account.

OpenAI: search recovery crosses midnight; GPT Image 2 edits recover separately

Dots, ChatGPT Work and some thinking experiences: search restored

OpenAI acknowledged search problems at 22:38 on 30 October and marked the incident resolved at 00:14 on 31 October. Its notice names Dots, ChatGPT Work and some thinking experiences. It does not say that every ChatGPT conversation or the whole API was unavailable. The resolution says affected services returned to normal. Official search incident.

The record supplies no exact impact onset, technical cause or separate mitigation timestamp. This search event is distinct from the persistent customer reports about existing Work conversations failing to load in our 30 October incident edition; it does not establish resolution of those accounts.

GPT Image 2 API: elevated errors on some image-edit requests

Some GPT Image 2 API edit requests had elevated errors on 30 October. OpenAI posted an identified status at 17:35, moved to monitoring after mitigation at 19:06, and posted resolution at 21:36. The affected component is Images in the API. Official image-edit incident.

The provider did not disclose the identified cause or establish the original impact start. This is a narrower finding than all image generation failing, and it does not establish an equivalent failure in ChatGPT’s consumer image interface.

ElevenCreative: image/video timeouts, stuck generations, Flows and MCP recovered

ElevenLabs acknowledged elevated latency and timeouts in ElevenCreative that could leave image and video generations stuck or failed. Its notice also names Flows and MCP. The investigation update is dated 30 October 12:04 Istanbul; the incident record itself was published at 12:19. Those timestamps describe different fields and should not be conflated. Original incident, official history.

The provider moved to monitoring at 13:49 and resolved the incident at 14:01. The notice does not establish an exact onset, cause, compensation policy or account-by-account recovery. Its ElevenCreative scope does not automatically extend to every music, speech or text-to-speech API product. It also does not resolve the separate sign-in and billing concerns in our earlier reporting.

Runway: delayed requests and failed submissions recover

Runway’s 30 October notice reports a disruption affecting its App, Backend and Public API components. At 22:16 Istanbul, Runway said it had identified the cause and was restoring service, without explaining the cause. At 22:21, it said a fix had been released, delayed requests should begin processing, and previously failed submissions could be retried. It marked the event resolved at 22:56. Official incident, original status API record.

The narrative acknowledges disruption even though the machine-readable record’s impact field is none and its component snapshots are operational. We retain that limitation rather than treating green flags as proof no incident occurred. Runway’s retry advice concerns failed submissions; it is not a direction to duplicate a request that is still pending. Exact user-impact boundaries remain unknown.

Grok: web, mobile, Imagine and Voice degradation resolved on 30 October

The official status history records degraded Grok performance affecting some users, followed by recovery. Investigation began at 04:45 Istanbul. The 04:52 update named Grok.com and Imagine; at 05:03, the notice also named Voice and the iOS and Android apps. At 05:29, most services had recovered, while Projects and Imagine remained degraded for some users. At 05:48, the provider reported normal performance across the named products. Original Grok notice.

This does not establish an outage of the general Grok API. The notice provides no technical root cause, affected-user count or exact onset before acknowledgement. The official page currently uses the SpaceXSI Status name and lists Grok and Cursor product groups separately.

Cursor and Grok Bot: six distinct workflow incidents, all resolved

The official history reports six additional events within the coverage window. The table preserves their distinct components and investigation-to-resolution notice times. These are operator notices for Cursor and Grok Bot workflows, not evidence that every upstream model provider was down.

The first 29 October record briefly labels a 01:40 recovery update resolved, then posts an investigating/monitoring message at 01:42. We use the later 02:13 final resolution and preserve the intermediate sequence. None of these notices establishes a shared cause across the six events.

Follow-up checks and coverage limits

Claude’s previously reported Calendar/Gmail recoveries and Google’s London postmortem had no material new update in the sources inspected. DeepSeek’s rendered history shows its latest October event on 27 October, before this edition’s overlap window. Claude history, Google Cloud history, DeepSeek history.

Original review and community checks did not establish a new corroborated customer incident or general resolution of the earlier Work and ElevenLabs access cases. Fresh isolated complaints remain unpublished leads. Positive ElevenLabs product/support feedback is also present; it does not resolve another customer’s account. Original ElevenLabs reviews.

Consumer Gemini history, tenant-only Microsoft notices and several specialist review feeds remain incomplete or inaccessible. Status pages can omit account-specific problems. This is a daily research snapshot, not continuous monitoring or an assurance that unlisted tools had no failures.

Sources

• OpenAI: search and GPT Image 2 edits.
• ElevenLabs: ElevenCreative incident and history.
• Runway: incident and updates.
• Grok and Cursor: official history, with individual incident permalinks in the sections above.

MUVO publishes this reporting and offers competing creative tools. See our reporting method and separate product-news edition.

Update timeline

• 31 October 2026, 11:55 Istanbul / 08:55 UTC: Research cutoff. Added eleven provider-confirmed resolved events, including previously inaccessible Grok/Cursor history inside the overlap window. Preserved unknown impact starts and distinguished ongoing customer-account concerns from these recoveries.