The lifecycle of a Timbra project
What to do before your event starts
A Timbra project moves through a series of stages, from the moment it's created on the dashboard to the point where captioning ends and any post-event exports are complete. This article walks through each stage and explains what to expect along the way.
1. Project creation
Create a new project from the Dashboard by selecting New project from… and choosing Live stream from HLS/RTMP/SRT. You can also import directly from a supported platform — for example, selecting Brightcove from the same menu will list your current live streams in Brightcove and allow you to import one.
Once created, the project is assigned the New status. This means it has been successfully set up and is ready to be configured, but captioning has not yet started.
2. Configuration
Before your event starts, use the Languages & settings tab on the project page to configure captioning and translation. Key settings include:
- Transcription language and provider — choose your primary (and optionally secondary) speech recognition service.
- Translation languages — Timbra supports translation into 250+ languages via Amazon, DeepL, and Google.
- Custom dictionaries and termbases — add speaker names, product names, and other domain-specific terms to improve transcription and translation quality.
- Audio track — by default Timbra uses the first audio track in the HLS stream; you can change this under Advanced options.
- Visibility — control who can see captions and the transcript, also under Advanced options.
For high-traffic events (more than 500 concurrent viewers or more than 2 concurrent streams), raise a support ticket at least 3 working days before your event so the necessary infrastructure can be confirmed. Make sure to include the following information:
- Expected number of viewers
- Number of concurrent streams
- Date and duration of the event
3. Stream status badges
Stream status badges appear at the bottom of the event project thumbnail on the dashboard and let you identify the current state of a project at a glance.
The possible statuses are:
New — the project has been created and is ready to configure.
Captioning — the event is live and captions are actively being generated.
Stopped — captioning was manually paused; the project can be restarted.
Ended — the source stream has ended and caption generation has stopped.
Source error — there is a problem with the incoming stream that is preventing Timbra from receiving it correctly.
Error — an issue was detected within the Timbra project itself. Open the project for more details.
4. Going live
Once your stream is active and you start captioning, the project moves to Captioning status. Timbra ingests the live stream and publishes captions in real time, synchronised with the audio.
For HLS-based projects, Timbra uses the stream's built-in buffer to produce captions just in time for viewers. For RTMP/SRT-based projects, audio and video are sent directly to Timbra.
During the event you can:
- Mute captioning to prevent music or pre-event content from being transcribed.
- Switch between transcription services manually under Advanced options if needed.
- Control the project from a mobile device by logging in to CaptionHub and opening the project.
- Control the project via the API for automated or integrated workflows.
5. Stopping captioning
Captioning can end in several ways:
Manually — you can stop captioning at any time from the project page. The project status changes to Stopped. You can restart captioning from a stopped project.
Automatically (HLS projects only) — HLS projects will stop automatically in any of these situations:
- The stream ends (no new video segments appear for 60 seconds)
- There are no viewers for 15 minutes
- No audio is received for 60 minutes
Note: SRT/RTMP-based projects do not stop automatically. They must be stopped manually.
When the source stream itself ends, the project status changes to Ended.
6. Post-event: export and reuse
After an event has finished, you can download a VTT subtitle file for every language from the History tab by clicking Download VTTs. These files are primarily intended for QA and auditing.
If you need captions for a VOD version of the event, we recommend creating a new CaptionHub project rather than reusing the live stream VTTs. A dedicated VOD workflow benefits from longer audio files (which improve speech recognition accuracy) and scene detection, which produces more natural caption splits. Linguists can also refine the captions in the CaptionHub editor before publishing.