AssemblyAI had 47 outages in the last 2 years totaling 34h 57m of downtime — averaging 1.9 incidents per month.
There were 47 AssemblyAI outages since December 19, 2024 totaling 34h 57m of downtime. Each is summarised below — incident details, duration, and resolution information.
From 8a - 10a GMT there was an issue related to the pick-up time of our non-English language models. This caused an increase in processing time for pre-recorded transcriptions. Our Engineering team has released a fix and all transcriptions should now be processing normally.
We are currently investigating an issue that is causing a small increase in turnaround times for STT services for pre-recorded audio for some requests. Traffic as a whole is still completing within the range of normal turnaround times. We will provide additional updates as we learn more.
Aside from a small window of about five minutes around 12:10pm PT where turnaround times were slightly higher than usual, we have continued to see turnaround times within the expected ranges so we are now closing this issue.
We experienced an issue with our Streaming API between 09:00 GMT and 10:04 GMT. This resulted in reduced quality of service for streaming sessions and some sessions were closing unexpectedly. A fix was implemented and our Streaming API is now functioning normally.
We are currently investigating an issue with Slam-1 that is preventing the successful completion of transcription requests using the model. We will share additional information as it becomes available. Please note this issue is only affecting requests using Slam-1 as the `speech_model`. Universal is unaffected.
We are continuing to investigate the issues affecting Slam-1. Currently, all requests using Slam-1 as the `speech_model` are failing with a "Server error, developers have been alerted" error. Our team is working to resolve this issue as quickly as possible. Note: This issue is only affecting requests using Slam-1 as the `speech_model`. Changing the `speech_model` to Universal will allow requests to complete successfully. See this section of our documentation for more information on specifying a speech model: https://www.assemblyai.com/docs/speech-to-text/pre-recorded-audio/select-the-speech-model
We have resolved the issue that was affecting Slam-1 and are now seeing requests being processed successfully. We will continue to monitor traffic to ensure good performance, but everything is working as expected at the moment.
We are once again experiencing issues with Slam-1 where requests are failing with a "Server error, developers have been alerted" error. Note: This issue is only affecting requests using Slam-1 as the `speech_model`. Changing the `speech_model` to Universal will allow requests to complete successfully. See this section of our documentation for more information on specifying a speech model: https://www.assemblyai.com/docs/speech-to-text/pre-recorded-audio/select-the-speech-model
We are seeing Slam-1 requests complete successfully once again, but we are still working to fully resolve the root cause of this issue.
We are still seeing good performance with the Slam model and requests completing as expected, but we are still working to fully resolve the root cause to prevent issues in the future.
We have continued to see good performance with no instances of increased errors over the last couple of hours ,but are continuing to monitor to ensure there are no further issues.
We have continued to see good performance for the last ~6 hours so we are marking this issue as resolved. To prevent future issues, we have made some improvements to allow Slam-1 to scale more effectively.
We are currently investigating an issue where there is a delay in new transcriptions showing up under the Transcription History tab of the dashboard.
Our team is still currently investigating the root cause of this delay.
The issue has been identified and a fix is being implemented.
A fix has been implemented and all customer dashboards should now be up to date. We are continuing to monitor the situation.
This incident has been resolved.
We had a brief period earlier where processing times for requests using the Slam-1 beta model took longer than expected. This slowdown occurred from approximately 7:37am PT to 7:57am PT and has now been resolved. Our team is continuing to monitor traffic to ensure continued good performance.
We are experiencing a formatting issue affecting English transcripts. Transcripts will continue to process successfully, but formatting may appear incorrect or incomplete. We're actively investigating this issue.
The English transcript formatting issue is now considered resolved. We are monitoring to ensure continued stability. Please re-run any affected transcripts to ensure correct formatting.
This incident has been resolved.
From 8:07pm to 8:22pm ET, transcriptions created with Slam-1 would have eventually resulted in a "Server error, developers have been alerted". Our team promptly looked into the issue and restored Slam-1 functionality.
We are currently investigating an issue that affects payment processing for accounts upgrading from the free tier to the pay as you go tier.
The issue has been identified and a fix is being implemented.
A fix has been implemented and we are monitoring the results.
This incident has been resolved.
We are currently experiencing issues that are affecting customers' abilities to add funds to their accounts. This is affecting some new and existing users. Our team is actively investigating the problem and working with our payment processor to resolve this as quickly as possible.
Our billing provider looks to have resolved the issues that were affecting them, and we are once again seeing payments complete as expected. We are monitoring the situation to ensure good performance going forward.
We have continued to see payment being processed normally so we are marking this issue as resolved.
We are currently investigating an issue causing degraded turnaround times for the Realtime API due to a slowdown in ASR processing. More updates to come as we learn more.
The issue has been identified and a fix is being implemented.
The system has been stabilized and we are continuing to monitor the results.
We have continued to see good results and are now marking this issue as resolved. We suspect that only a small number of transcripts were affected as P50 and P90 response times were within reasonable thresholds throughout, with only P99 response times being affected initially.
We are currently investigating an intermittent issue with downloading audio files.
The issue has been identified and a fix is being worked on. Only a few customers have been impacted.
A fix has been implemented and we are monitoring the results. We observed that errors stopped at 9AM EST.
This incident has been resolved.
We are currently investigating an issue that is preventing the retrieval of async and LeMUR transcriptions on the US and EU endpoint.
We are continuing to investigate this issue of a degraded performance of our async API and LeMUR to retrieve transcripts.
The issue has been identified and a fix is being implemented. After investigating, this seems to be affecting a small subset of transcripts.
We are continuing to work on a fix for this issue.
A fix has been implemented and we are monitoring the results.
We are continuing to monitor for any further issues.
This incident has been resolved.
We are currently investigating an increase in authentication errors. These errors typically include the following error message: `Authentication error, API token missing/invalid` If you get this error please retry the request as retries are typically successful.
We are continuing to investigate the cause of the increased authentication errors but this issue is only affecting a small number of requests relatative to our overall traffic. We are continuing to see over a 99.9% success rate for requests to our APIs.
We have seen a reduction in authentication errors over the last ~60 minutes but we are continuing to monitor traffic to ensure continued good performance.
We have continued to see good results over the last several hours and are now marking this issue as resolved.
We experienced a customer-facing incident that impacted transcription turnaround times for our async endpoint.
We are currently working with our billing provider to address an issue that is causing certain dashboard information to not be updated as expected. This issue includes usage charts, spend charts, and balance info not reflecting current values. This issue should have no impact on API usage.
Our billing provider has identified the root cause of this issue and is working to resolve it.
Usage data is once again starting to flow through the system, but the information backlog will take time to be completely processed. The current estimate is that it will take 1-1.5 hours to process this backlog. We will provide another update once all information has been processed.
This issue has been resolved and all dashboard information is now properly updating.
On Tuesday, February 25th from 1:01 AM to 1:08 AM Eastern Time, our service experienced a brief disruption during a routine database maintenance. During this 7-minute window, some API requests returned 500 errors, while other transcription requests were temporarily delayed in our processing queue. Any transcripts successfully queued were automatically retried and processed by our systems. Our team quickly identified the issue, reverted the changes, and fully restored service functionality by 1:08 AM. We continue to actively monitor our Async API to ensure everything is operating normally.
We are currently investigating an issue that is causing account information to not update correctly. This can affect some of the information displayed in the dashboard including account balance, which in some edge cases involving recent payments can affect API usage.
We have released a fix aimed to address the account issues we have been experiencing and are monitoring to ensure that the issue has been resolved.
We have continued to see good performance and previously affected users are reporting that the issues are resolved on their end as well so we are marking this issue as resolved.
We are investigating an issue that is causing a minority of requests to our Async transcription to not complete.
We are experiencing a service degradation affecting the completion rate of English transcription requests through our Asynchronous API.
We have identified that the service degradation is specifically affecting English transcription requests for audio files between 25 and 600 minutes in duration. Our engineering team is actively deploying a fix.
A fix has been deployed and we are monitoring results.
We are continuing to monitor for any further issues.
All previously queued transcription requests have now completed processing. New requests are being processed within standard response times. Please retry any requests that failed during this incident. Service has been fully restored.
We have identified an issue in which users are unable to access the API after upgrading their account without auto-refill enabled. Our team has a fix in place for this issue and is working to deploy it to Production.
We are continuing to work through the issue affecting some accounts' ability to use the API. This issue is only affecting a very small subset of customers who have recently created or upgraded their accounts. If you are affected by this issue please reach out to our Support team at [email protected] and we can help unlock your API access.
Our team has identified a cause of this issue and is currently working to deploy a fix.
We have released a fix to address the issues that were affecting a small subset of new or upgraded accounts. These issues now look to be resolved but we will continue to monitor traffic to ensure continued good performance.
We have continued to see good performance since the fix was released so we are marking this issue as resolved.
We have identified a configuration issue in our deployment setup that caused requests to our streaming service to return 503 errors. We have released a hotfix for this issue and requests to our streaming service are now completing successfully. We will continue to monitor results to ensure the service is working as intended.
We have continued to see requests to our Streaming service complete successfully, so this issue is now being marked as resolved.