v2.5.0

Recordings, audio, transcripts and logs can now be written to your own S3-compatible bucket instead of ours, configured per team. Transcription also retries a provider that fails transiently rather than giving up on the first error.

✨New Features

• Bring your own storage. PUT /v2/storage-config sets an S3-compatible destination for your team — custom endpoint, region, path-style addressing and separate buckets for artifacts, audio and logs — and POST /v2/storage-config/test verifies the credentials before you commit to it. GET and DELETE read and remove it, and there is a Storage tab in the dashboard. Recordings, audio chunks, transcripts, speaker timelines, screenshots and logs all follow the configured destination, and signed URLs in webhook payloads point at your bucket.
• Teams with no storage configuration continue on Meeting BaaS storage exactly as before.

⚡Improvements

• A transcription submission that fails for a transient reason — a timeout, a network error, a rate limit or a provider 5xx — is retried with backoff. Authentication and validation errors still fail immediately, since retrying them cannot help.
• A superseded transcription attempt can no longer overwrite the result of a newer one.

↩ All releases