Transcribe and index a folder of videos

Add a folder to Auto jobs

1. Prepare a local folder of supported media and grant Bonframe access to it.
2. Open Auto jobs or the folder-indexing view and drop the folder, or use the folder-selection control.
3. Review the files found and add them to the queue. Existing work can continue while the queue records the remaining files.
4. Watch queued, running, completed and needs-attention states. Check transcription and index readiness before searching a project.

What a batch index job produces

The folder workflow can prepare speech recognition, a text index and a frame index for each video. These results make the library searchable; they are not automatically translated, dubbed or lip-synced outputs.

Speaker information and recognition quality depend on the selected model and recording. Review important names, numbers and speaker labels before using a transcript for production.

Pause, resume and retry deliberately

• Pause stops new queue work at the supported boundary; an active native operation may need time to finish.
• Resume continues queued work. Check whether the same source already has a job before adding the folder again.
• For a failed item, read the reported stage and error. Fix file access, a missing model or device resources before retrying.
• Cancellation is not a rollback of files and results already produced.

Choose a transcription model

Use the model choices and recommendations currently shown by the app for your device and source language. Test a short section with representative accents, noise and speaker changes. Do not select a model solely from an estimated minutes-per-hour table.

For a large folder, verify the first few results before allowing the entire collection to run unattended.

Related guides

• Devices and local model setup
• Search your video library with Memory
• Task history, queues and troubleshooting