Deepgram is up and running. Deepgram provides speech-to-text and voice AI APIs.
Worked out from the incidents Deepgram reported, so it is an estimate and not a measurement.
Last incident: “Elevated 5xx Errors”, today.
We have identified the issue & are monitoring the solution.
We are continuing to investigate this issue.
We are currently investigating elevated 5xx errors with authentication services at one of our datacenters.
This issue has been resolved.
This issue has been resolved.
We experienced 2 bursts of elevated error rates between 02:58 - 03:16 UTC and 03:53 - 04:18 UTC. The issue is now resolved.
This incident has been resolved.
This incident has been resolved.
Deepgram’s official status page says: All Systems Operational. This page checks it again when you load it, at most every 5 minutes.
The most recent incident Deepgram reported was “Elevated 5xx Errors”, today (2026-10-22 16:00 UTC). None of its incidents in the last 90 days were a full outage.
Deepgram reports 13 components on its status page, including Batch API, Developer Documentation, License Validation, Streaming API, Usage API. Each is checked again whenever this page refreshes.
Retry with exponential backoff, fall back to another model or provider if your product allows it, and avoid hammering the API while it recovers.
Deepgram tracks the Streaming API and the Batch API as separate components — a WebSocket/streaming incident doesn't stop already-queued batch jobs.
They're separate components (TTS API vs the transcription APIs and the Voice Agent API). Check the TTS row rather than the public API row.
If the status above is operational but Deepgram still fails for you, the cause is probably on your side: try another network, clear the cache, or check your own connection. If it shows a degradation or an outage, others are affected too.
OpenUptime checks your site every minute and sends an alert to Slack, Telegram, email or a webhook the moment it stops responding. It is open source and free to start.