Is Phonely Down?

Voice agent response times have been stable since the capacity changes we made on November 9, and voice failovers have stayed at low levels since then. This incident is resolved. We apologize for the disruption to affected calls. --- [2026-11-10T17:31:10Z] (monitoring) Between 14:10 and 21:00 UTC on November 9, some voice agent calls had a delay of a few seconds before the agent responded. A small share of calls ended early as a result. The cause was a capacity constraint in one of the systems that generates agent responses, not in our voice providers. We have added capacity and moved that system to new infrastructure, and response times have been back within normal range since late November 9. We are continuing to monitor and are working on a further change so that slow responses no longer interrupt audio. We will post an update when this incident is resolved. --- [2026-11-09T19:28:50Z] (investigating) We are investigating an increase in voice failovers affecting some calls across multiple voice providers. Fallback connections are functioning as expected, but some callers may experience delays in voice responses. The team is investigating the cause and monitoring the impact.

This incident has been resolved. --- [2026-10-30T00:37:22Z] (monitoring) We’ve applied a fix for the issue where the AI agent rarely didn’t speak on calls. The cause was one of our internal AI models slowing down during periods of high call volume. We’ve moved affected traffic off that model, and call behavior is returning to normal. We’re monitoring closely and will post a final update once we’ve confirmed the issue is fully resolved. --- [2026-10-30T18:03:16Z] (identified) We believe we’ve identified the cause of calls where the AI agent doesn’t speak. It’s linked to one of our internal AI models slowing down during periods of high call volume, which delayed the agent’s responses long enough that callers heard silence. We’re now applying a fix and will share another update shortly. --- [2026-10-30T18:02:08Z] (investigating) We’re investigating an issue where some calls connect but the AI agent doesn’t speak, so callers hear silence and often hang up. This mainly affects outbound calls on Enterprise, with a smaller number of inbound and Production calls also affected. It has been worse during high-volume calling periods. Most calls are still connecting and completing normally. We’ve identified where in the call the audio fails and are working to find and fix the cause. Next update within 2 hours.

Agent emails, including post-call emails, have been sending normally since we deployed our changes yesterday afternoon. The failed emails were caused by connections from our servers to our email service that sometimes stalled before an email could be sent. We now retry those connections automatically, and we changed how that traffic is routed. We'll continue to watch email delivery closely. If you're missing an email from October 28 or 29, please contact us and we'll help. --- [2026-10-30T15:43:13Z] (monitoring) All emails from yesterday that failed to send have been sent out as of 4:05 PST on 10/29/2026 --- [2026-10-30T14:16:13Z] (monitoring) Since we deployed our change yesterday afternoon, agent emails, including post-call emails, have been sending without failures. We're continuing to watch email delivery closely through today's busiest hours before we close this incident. If you're missing an email from October 28 or 29, please contact us and we'll help. --- [2026-10-30T00:03:01Z] (monitoring) Our fixes for the intermittent email send failures are in place, and we have not seen any further failures since they were deployed. We are continuing to monitor and will post another update once we confirm the issue is resolved. --- [2026-10-29T19:58:01Z] (identified) We've deployed a change to reduce the intermittent failures sending agent emails, including post-call emails. We haven't yet confirmed that it resolves the issue. At times, our servers have stalled while opening a secure connection to our email service, and those emails failed to send. Our servers now give up on a stalled connection after a few seconds and retry it automatically. We have also simplified the network path these requests take. We're still investigating why the connections stall. We're watching email delivery closely and will post another update once we know whether these changes have stopped the failures. If you're missing an email from the last two days, please contact us and we'll help. --- [2026-10-29T19:14:57Z] (investigating) We have identified the cause of an issue where some emails sent by agents, including post-call emails, failed to send for a small number of customers. The failures come from intermittent connection timeouts between our platform and our email provider. We are deploying a fix that automatically retries these connections, and we are reviewing affected emails so they can be resent. We will post another update once the fix is in place. --- [2026-10-29T18:23:14Z] (investigating) We are aware of an issue impacting a small number of customers where emails are failing to send. We are investigating.