The issue has been fixed; this affected only newly created accounts today. Impact is less than 100 workspaces.

The issue has been resolved, and all the voices should be visible now. This only affected workspaces with 50,000 or more voices.

The root cause was an incompatibility in how ElevenLabs handled mid-dialog re-INVITEs. When SIP providers sent re-INVITEs, ElevenLabs responded with 488, which some providers treat as fatal, issue a BYE, and terminate the call. Per RFC 3261, a 488 should not necessarily result in session termination, but provider behaviour varied.

The issue has been fully resolved. The issue was only with the calls on the EU Residency.

We were able to identify the root cause of the increase in errors. We have implemented a fix to mitigate any possible occurrences moving forward.

EU Residency:

Starting Saturday, March 14th at approximately 17:00 UTC, some calls were affected by connection failures, or connected with silent audio. This was fully resolved at 08:00 UTC on March 16th.

Cross-Region:

Starting Sunday, March 15th at approximately 01:00 UTC, some conversations experienced a 1 second delay to initial audio. In some cases, this caused disconnections due to how some SIP trunks handled the audio delay. This issue was fully resolved at 16:00 UTC on March 16th.

Our team has resolved the issue, services are back to normal.

We have confirmed full functionality of both EU and IN residency webpages.

We have confirmed with affected clients and in our analytics that SIP trunking functionality is now fully restored.

We have confirmed fully restored functionality of the voice of "George".

The issue has been resolved.

Gemini model performance has fully recovered; Our cloud provider has informed us that the upstream issue is fully resolved, and we have carefully tracked our metrics to confirm this.

Error rates have returned to the baseline.

As of 7:45 PM UTC, all affected services have fully recovered and are operating normally.

Agents Platform and Audio Platform (Website) have returned to full health. Text-to-Speech (TTS), Speech-to-Text (STT), and Dubbing APIs were unaffected throughout the incident.

The root cause has been identified and remediated. We will be conducting a thorough post-incident review and will share further details in a follow-up report.

We apologize for the disruption and appreciate your patience throughout this incident. If you are still experiencing any issues, please don't hesitate to contact our support team.

The issue has been resolved.

We had a temporary outage between 13:26 and 13:56 UTC, which resulted in a number of post call webhook failures. This issue should now be resolved.

The error rates have now returned to normal levels

The RAG performance issue is now resolved.

The fix to resolve the missing conversation_id in Post Call Transcripts has been successfully deployed. A detailed RCA will be published later.

We have confirmed the release of our Data Extraction fix has continued to mitigate issue occurrence during peak traffic and that rates of Data Extraction failures are below pre-incident baseline.

The issue with Static SIP traffic on the EU residency server has been fully resolved.

The issue has been fully resolved.

LLM error rates have normalized, and Gemini-2.5-Flash conversation latency is back within expected ranges. We’ll continue monitoring closely. Improvements accommodating reduced cloud provider have been implemented, and will be live shortly.

We have confirmed that our fix successfully remediated the issue, and Dub Exportation is fully functional. Previous download attempts may need to be resubmitted.

The issue with RAG indexing has been resolved.

The issue has been identified and fixed.

The issue has been fully resolved.

Error rates have stabilized. We are continuing to enhance LLM fallback coverage to improve system resilience and reduce the risk of single points of failure.

The issue has been completely resolved, and the requests with the v3 model are being processed.

We’ve resolved the issue impacting our support intake.

Following a surge of automated spam affecting our support workflows, we have chosen to disable our public support submission forms.

Support remains readily available via our support email, team@elevenlabs.io.

We are taking this step due to repeated spam-related disruptions and insufficient mitigation from our support platform provider. Please note that ElevenLabs services have not been affected in any way.

This issue is now fully resolved.

A fix has been deployed to prevent inflated per-request charges affecting a subset of Speech-to-Text requests. The issue has been fully resolved.

We are calculating the impacted usage during the incident window and will refund the missing credits to all affected accounts to correct the overcharging.

The issue has been fully resolved.

The issue related to transcription for EU residency server on the Agents platform has now been resolved.

Remediation measures have been fully rolled out, restoring failure rates to baseline. The root cause has been addressed, and the issue is now fully resolved.

Between 15:38-17:14 UTC, January 5th, less than 0.1 percent of Agent Platform conversations failed due to "Unexpected server error" messages. This issue is now addressed and systems are completely functional.