Transcription, AI models, payments & developer workflows · Daily incidents

AI incidents: ElevenLabs Scribe and Claude recover, OpenAI payment issue resolves — 11 November 2026

ElevenLabs Scribe v3 Live and two Claude models recovered. OpenAI resolved Mastercard declines; Grok usage data, Fireworks and Bolt also had incidents.

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.

Scribe and Claude recovered; an older payment incident is closed

ElevenLabs says Scribe v3 Live recovered, Anthropic reports a brief resolved error period for Fable 5.2 and Mythos 5.2, and OpenAI has resolved its earlier U.S. business Mastercard payment issue. The notices concern specific models and services. They do not establish a single shared outage.

This edition covers 8 November at 12:05 through 11 November at 11:53 Europe/Istanbul (UTC+3), including the 48-hour overlap from our previous successful scan. Previously covered incidents appear only for material updates. Times below are Istanbul time unless another timezone is explicitly stated. Provider resolution is not proof that every failed request, payment or job subsequently completed.

Service / status Notice times, Istanbul
ElevenLabs Scribe v3 Live — resolved 11 Nov: acknowledged 04:10; recovery/monitoring 04:19; resolved 06:03
Claude Fable 5.2 / Mythos 5.2 — resolved 11 Nov: retrospective notice 00:47; provider says four minutes of errors
OpenAI business Mastercard payments — resolved continuation 10 Nov: resolved 19:06, after monitoring since 7 Nov
Grok usage dashboard / APIs — resolved 10 Nov: acknowledged 21:32; resolved 22:35
Fireworks DeepSeek V4.1 Flash 1017 — resolved 11 Nov: acknowledged 06:45; resolved 08:10
GitHub Git / pull-request workflows — resolved 10 Nov: acknowledged 19:10; resolved 20:55
Bolt Prompting — two resolved notices 11 Nov: 08:22–08:27 and 08:29–08:34

These are notice milestones, except where a provider explicitly describes an affected interval below. Unknown impact boundaries stay unknown.

ElevenLabs: Scribe v3 Live is available again

ElevenLabs reported Scribe v3 Live unavailable, then confirmed recovery. Its Global (United States) notice tags the ElevenAPI Speech to Text component. Investigation was posted at 04:10 on 11 November, recovery and monitoring at 04:19, and final resolution at 06:03. This does not establish an outage of ElevenLabs music, voice generation or every speech service. Regional incident notice.

The provider describes a 21-minute outage, giving the boundaries as 00:50–01:11 “BST.” We retain that wording because the timezone label is ambiguous in this November notice; we have not converted those impact boundaries into Istanbul time. The Istanbul milestones above use the notice’s UTC publication timestamps. Its longer open/monitoring interval must not be presented as user downtime. No technical cause or user workaround was supplied.

Claude: a short Fable 5.2 and Mythos 5.2 error period

Anthropic reports four minutes of overloaded errors on Fable 5.2 and Mythos 5.2, already mitigated when it posted the notice. The resolved update was published at 00:47 on 11 November Istanbul, or 21:47 UTC on 10 November. The four-minute duration is Anthropic’s statement; exact start and recovery times are absent. Claude incident.

The notice does not specify plans, regions, a technical cause or affected application/API surfaces. It therefore cannot support a claim that all Claude users or models were unavailable. This is separate from the earlier Fable 5.1 recovery.

OpenAI: earlier Mastercard payment declines are resolved

OpenAI closed its U.S. business Mastercard incident at 19:06 on 10 November. The issue could prevent API automatic recharges and one-time credit purchases for specified cards; exhausted prepaid balances could then cause API calls to fail. Monitoring had begun at 04:29 on 7 November, following the first acknowledgment at 21:38 on 6 November. These notice dates do not establish continuous impact throughout that period. OpenAI payment notice.

We have added the resolution to the original 7 November report, preserving its publication date and earlier timeline. This is a recovery update, not a new payment outage. OpenAI’s earlier advice was to use another issuer’s card for affected payments; it did not state that every previously declined purchase would automatically complete. No detailed cause was disclosed.

Grok: usage data was slow or missing for some users

xAI reports that usage information on its dashboard and usage APIs is loading normally again. Its Grok-component notice opened at 21:32 on 10 November, describing slow data or no displayed usage for some users, and resolved at 22:35. It did not say that model responses stopped or that invoices were incorrect. Actual impact start, cause and a workaround were not supplied. Grok usage-data notice, original incident feed.

Fireworks: DeepSeek hosting recovers; an older Kimi notice is unavailable

Fireworks resolved degraded performance for its hosted DeepSeek V4.1 Flash 1017 component at 08:10 on 11 November. Its component record runs from 06:45:26 to 08:10:52, approximately 85 minutes. That measures the provider-recorded degraded interval; earliest user impact is not independently established. No cause or workaround was given. This is a Fireworks hosting incident, not evidence of a disruption across DeepSeek’s own service. Fireworks incident.

The separate Kimi K3 notice we reported yesterday no longer exposes its incident details. It also does not appear in the retrieved history. It was investigating at yesterday’s cutoff; its present resolution time and status are unverified. We have updated yesterday’s report to make that limitation explicit. Disappearance of a notice does not establish recovery or continuous failure. The separate Kimi K3 US recovery remains documented. Previously accessible Kimi notice, Fireworks history.

Fireworks’ older deployment, training and batch-inference notice still carries its 6 November restoration/monitoring update, with no fresh resolution established.

GitHub and Bolt: separate developer-workflow disruptions

GitHub resolved elevated response times for Git operations and pull requests at 20:55 on 10 November. The first acknowledgment was at 19:10. Later updates included delayed Actions workflows triggered by pull requests, delays and failures in pull-request processing, and partial recovery. Actions was marked operational at 20:28, before the final incident resolution. A detailed root-cause analysis was promised but absent at cutoff. This workflow incident does not establish a Copilot-model outage. GitHub notice.

Bolt recorded two separate Prompting degradations on 11 November, with component intervals 08:22:38–08:27:44 and 08:29:14–08:34:31. Both are resolved. The notices give no cause, broader impact or workaround; they should not be merged into one uninterrupted outage. First Bolt notice, second Bolt notice.

Sources

Original notices are cited beside each finding. We rechecked Google AI Studio’s rendered history, Google Cloud, Google Workspace, the wider creative-tool watchlist and accessible original customer reviews. No additional qualifying Google incident was established. Isolated account complaints and speculative community predictions remain unpublished leads; blocked review pages and incomplete histories limit coverage. This is a daily scan, not a claim that every service or account is healthy. See our reporting method.

Update timeline