Changelog

August 11, 2026

Meeting Scribe speaks your language.

Meeting Scribe speaks your language.

Meeting Scribe now detects the language being spoken on your incident call and transcribes it automatically. Join the bridge and talk in the language your team actually works in: the Scribe will identify the language and start transcribing automatically.

The transcript does more than sit in a file. Meeting Scribe works together with Rootly Agents, like the @Rootly Agent in Slack, feeding it everything said on the call as live conversation context. Teammates who never joined the call can mention @Rootly in the incident channel and catch up asynchronously—what happened, what has been tried, and what was decided—no matter which language was spoken on the bridge.

What's included

  • Automatic detection of the spoken language on every recorded incident call
  • Automatic transcription of the call in the detected language
  • Transcripts feed the Rootly Agents as real-time conversation context, alongside your incident data and channel discussion
  • Anyone on the team can gather context async by asking the agent, regardless of the language spoken on the call

Why it matters

Virtual meetings are where the best and most efficient collaboration happens. Theories are discussed, decisions get made, and tasks are completed. When the Scribe captures the call in any language and hands it to the AI Agent, everyone stays aligned and up to date on the incident—each person has what they need to contribute, and incidents get mitigated and resolved faster.

After the incident is resolved, your teams are able to revisit the transcript and pull out key learnings to drive reliability long-term.

Getting started

See the Meeting Scribe documentation to learn more.

What else shipped

Improvements

On-Call

  • You can now select multiple acknowledged alerts and snooze them all at once.
  • SMS alert notifications now include the alert's title and description so you get more context at a glance.
  • Alert deduplication now works over a longer time window instead of being capped at alerts from the last 31 days.
  • New user profiles now default to email-only for quiet notifications.
  • SMS notifications can now be delivered to phone numbers in Sri Lanka.

Incident Response

  • Heartbeats can now be scoped to a specific team, so only that team's members can view and edit them.
  • Slack channel pickers now only show channels you are a member of, so you cannot target channels you should not see.
  • You can now use variables, like the incident's severity, to set the priority when a workflow creates a Linear ticket.
  • Workflows that create Zoom meetings can now start recording without waiting for the host to join.
  • You can now turn off the automatically added Follow Up Items section on retrospective exports.
  • Retrospectives now have a single PDF export option instead of two duplicate buttons.
API

On-Call

  • A new GET /v1/oncalls endpoint lets you ask who would be paged for a service right now.
  • The GET /v1/schedules endpoint can now filter schedules by team, matching the filter in the web UI.
  • The escalation policies API now supports filtering by team.

Incident Response

  • GET /v1/incidents now returns a 400 error for badly formed date filters instead of quietly returning too many results.
  • The API playground now shows array fields, so you can add list values when trying out endpoints.
  • Cleaned up the status page API schemas so the public API reference docs validate and display correctly.
Terraform
  • Email alert sources now return their settings correctly through the API, so accept_threaded_emails no longer causes a permanent Terraform diff.
  • Clearing a service's or functionality's owning teams by sending an empty list now works instead of being silently ignored.
Fixes

On-Call

  • Callers now hear a message telling them to hold for the next person when the first responder is unavailable during live call routing.
  • Slack alert notifications now show timestamps in the matched escalation path's own timezone when one is set.
  • Filters you add on top of a saved alert filter now stay applied when you run a text search.
  • The source filter on the alerts page now lists all of your alert sources on first load, without needing a refresh.
  • The Terraform export for webhook alert sources now produces valid configuration code for field mappings.

Incident Response

  • SLA item numbers now update right away when you reorder the list.
  • Custom sort order for services and functionalities in the catalog now saves correctly after a page reload.
  • Corrected the wording on the "Start New Session" button in the meeting recordings screen.

You and your teams deserve
modern incident management.

Get a 1:1 demo with one of our technical staff or start your free 14-day trial.