Is Boost AI down?

We are currently observing increased processing times for requests to the OpenAI model text-embedding-3-large in the US region. Requests may consequently take longer than usual. Our team is investigating the cause. We will provide an update as soon as more information is available.

We are currently seeing improvements and increased stability across affected environments following the earlier disruption impacting the OpenAI model. Processing times have returned toward normal levels, and most requests are now processing with normal processing time. We are continuing to closely monitor system performance.

This incident has been resolved.

On 11 August 2026, from 14:30 to 19:50 CEST, a subset of virtual agent instances experienced degraded performance on conversations processed via the OpenAI (via boost.ai) vendor configuration, caused by elevated latency on the upstream provider's models. During this window, affected conversations experienced extended processing times and a elevated number of timeouts or errors. Provider latency returned to normal at 19:50 CEST, after which all affected instances were stable. Total duration of degraded performance: 5 hours 20 minutes. Note: this incident entry was published retroactively on the 20th of August as part of our commitment to a complete and accurate public service record.

On 6 August 2026, from 15:28 to 18:53 CEST, a subset of virtual agent instances experienced degraded performance on conversations processed via the OpenAI (via boost.ai) vendor configuration, caused by elevated latency on the upstream provider's models. During this window, affected conversations experienced extended processing times and a elevated number of timeouts or errors. Provider latency returned to normal at 18:53 CEST, after which all affected instances were stable. Total duration of degraded performance: approximately 3 hours 25 minutes. Note: this incident entry was published retroactively on the 20th of August as part of our commitment to a complete and accurate public service record.

On 28 July 2026, from 13:03 to 19:50 CEST, a subset of virtual agent instances experienced degraded performance on conversations processed via the OpenAI (via boost.ai) vendor configuration, caused by elevated latency on the upstream provider's models. During this window, affected conversations experienced extended processing times and a elevated number of timeouts or errors. Provider latency returned to normal at 19:50 CEST, after which all affected instances were stable. Total duration of degraded performance: approximately 6 hours 47 minutes. Note: this incident entry was published retroactively on the 20th of August as part of our commitment to a complete and accurate public service record.