AI-Media Outage History

On September 23, 2026, some existing jobs using Falcon Classic for legacy HTTP POST workflows stopped delivering captions due to an issue introduced during a planned platform update. Other Falcon workflows and services were not affected. Impact: Some Falcon Classic jobs that were already running were interrupted between approximately 1:54 PM ET and 6:45 PM ET. Affected jobs may have shown as running but were not delivering captions. New jobs started between 2:19 PM ET and 3:11 PM ET were also at risk. Impact to new jobs was minimal. Resolution: All affected jobs were moved back to a stable environment, and caption delivery has been fully restored. We are reviewing our update process to prevent this from happening again. If you are still experiencing issues with a Falcon Classic job, please contact support.

At approximately 3:15 PM EDT on July 29, an AWS host-level failure caused one Lexi worker instance to be replaced automatically. Because the instance was terminated forcefully, jobs running on it could not be requeued and were left in a state where they appeared active but were no longer producing captions. 16 jobs were affected, and captioning resumed on restart. All affected jobs have been cleared, and Lexi is operating normally.

At 06:56 AM EDT on July 24, an AWS routing issue in us-west-2 caused several client connections to drop across two US relays within a three minute window. All affected clients reconnected cleanly to other US relays. Sydney relays were unaffected. Connection stability returned to normal following the reconnects.

On July 24, an expired backend credential left the iCap Connection Stability and iCap Server Cluster Reconnects metrics blank for approximately 15 hours. This overlapped with the relay disconnects reported separately, so customers checking these metrics that morning had no visibility into connection health. Metrics for this window show a gap.

US East has maintained normal service since the fix was deployed and we are marking this as resolved. We are continuing to monitor and ensuring robust solutions are in place to continue consistent performance. Please reach out to support if you encounter any issues. Thank you.

We have seen successful connections return to Ai-Live LEXI from 10:45am AEST / 6:45pm EST in the AU region as services resume. If you encounter any persisting issues, please contact support.

Date: November 7, 2025, approximately 7:00 PM - 7:03 PM ET Status: Resolved Summary: We experienced a brief service disruption to LEXI Live due to a sudden traffic spike. Some users experienced delayed connections or temporary disconnections. Impact: New transcription jobs experienced delays of up to 2-5 minutes in establishing connections Some active transcription sessions experienced brief disconnections (ranging from 48 seconds to 2+ minutes) Root Cause: A rapid spike in traffic exceeded our standard capacity, causing a temporary backlog while our autoscaling systems provisioned additional capacity. Resolution: Service automatically recovered as additional capacity came online. All systems returned to normal operations shortly after. Next Steps: We are implementing several improvements to our auto-scaling infrastructure. These include faster response times to traffic spikes and increased baseline capacity to better handle peak usage periods.

The Ai-Live US instance experienced an outage at just before 11:00am EST at which time clients reported the Ai-Live Viewer link was down. The issue was caused by unexpected Viewer load in multiple Ai-Live sessions. Ai-Live server resources were added at 11:35am EST in order to accommodate the increased load. Additional resources were enabled at 3:51 and 3:57pm EST to mitigate increasing CPU utilization as viewership continued to grow. The issue was fully resolved at 4:00pm EST.

Lexi services have stabilized and we are comfortable marking this incident as resolved. All systems are operating normally following the AWS outage. If you experience any issues, please contact support.

One of our iCap server nodes carrying iCap and Lexi traffic became unresponsive and required manual intervention to restore. Other servers should have been able to fulfill most jobs but some temporary interruptions occurred due to transfers, retries, and related conditions. Issues began about 02:35 UTC and were clear at 03:15 UTC.