AI incidents: OpenAI API fix, ChatGPT Work and ElevenLabs access reports — 30 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 Support has explained a Conversations API fix; fresh users still report ChatGPT Work conversations failing to load and ElevenLabs account-access problems. The API finding is an attributed provider response. The access findings are corroborated customer accounts, not confirmed company-wide outages.

This edition covers 27 October at 12:08 through 30 October at 12:06 Europe/Istanbul (UTC+3). It includes newly published explanations of older failures and updates to unresolved reports. Events did not necessarily begin on the edition date.

OpenAI explains a Conversations API fix after intermittent 503 errors

On 29 October at 20:15 Istanbul, OpenAI Support said two reported Conversations API failures reached servers restarting after exhausting memory. It said a 26 October change to reading long conversation histories, plus additional memory, stopped those restarts. The response supplies no exact fix time, impact end time or complete affected population. Official support reply.

Separate the endpoint correction from the original thread title

The discussion began as a suspected Responses API problem. On 26 October at 16:25 Istanbul, the original reporter clarified that disabling automatic Conversations API creation stopped their errors while Responses remained in use. On 27 October at 21:43, OpenAI Support suggested chaining Responses calls with previous_response_id to retain context without a separate Conversations call. Reporter’s correction, support’s workaround.

The initial reported onset, 26 October at 01:00 Istanbul / 25 October at 22:00 UTC, is approximate and customer-supplied. It is not an official incident start. We therefore do not calculate an outage duration. Support’s recovery statement concerns the traced restarts, not proof that every affected workflow recovered.

A separate contributor used Azure OpenAI in UK South and later reported their Responses-with-tools problem no longer occurring. OpenAI directed that contributor to Azure Support; its Conversations explanation does not establish the cause of an Azure incident. Separate customer follow-up, 28 October.

ChatGPT Work conversation-loading failures continue in fresh accounts

Two independent users reported loading problems on 29 October in an existing ChatGPT Work thread. At 12:49 Istanbul, one said some conversations still would not load after updating. At 17:57, another described inaccessible project chats while a newly created chat could open. First account, second account.

This continues the issue covered in our 21 October incident edition. It is not evidence of a new global outage or proof that conversation data was deleted. A fresh chat opening is a limited observation, not a verified way to recover project history. We found no provider-confirmed cause, impact interval or general resolution for these latest accounts.

ElevenLabs: two dated reviews describe account-access problems

Two independent reviews describe difficulty accessing ElevenLabs accounts, with experience dates of 27 and 29 October. The 29 October account describes verification and password-reset loops, temporary access, and renewed sign-in failure. The 27 October account describes being locked out and support exchanges that had not resolved access. 29 October review, 27 October review.

Status: customer-reported and unresolved at the level we can verify. These accounts do not prove one shared technical cause, prevalence or an ElevenLabs API outage. Exact onset, recovery and publication times are unknown; the experience dates are not outage timestamps. No account-specific provider resolution was visible. Separate positive product feedback does not resolve these cases. The complete entries were inspected in the original listing; individual permalink retrieval was restricted.

Follow-up checks and coverage limits

Previously reported Claude Calendar recovery and Google’s London postmortem had no material new update in the sources inspected. OpenAI’s public status history did not separately establish the scope of the Conversations issue; the support reply above is the relevant original evidence. Runway’s earlier Magnific recovery remained unchanged. Claude incident history, OpenAI history, Google Cloud history, Runway history.

ElevenLabs’ accessible history supplied no fresh provider acknowledgement of these sign-in reports. Its earlier billing concerns remain a separate issue in the 29 October edition. ElevenLabs history.

Consumer Gemini history, tenant-specific Microsoft notices and several specialist review feeds remain incomplete or inaccessible. A green current-status indicator cannot establish that no earlier disruption occurred. Isolated IP-blocking, credit or quality complaints were not combined into the access finding.

Sources

• OpenAI: support’s Conversations API explanation and reporter’s endpoint correction.
• ChatGPT Work: first dated loading account and independent project-chat account.
• ElevenLabs: original review listing; review-specific links appear above.

Original support messages, dated firsthand accounts and provider histories are linked beside the relevant claims. Community contributors are not necessarily verified paying customers. We have not accessed their accounts or independently tested the reported failures.

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

Update timeline

• 30 October 2026, 12:06 Istanbul / 09:06 UTC: Research cutoff. Added OpenAI’s newly published explanation of the 26 October Conversations fix, fresh ChatGPT Work loading accounts and corroborated ElevenLabs access concerns. Unknown event times remain unknown.