CuevueHelp Center

Transcription & Search

Cuevue converts uploaded audio and video into a timestamped transcript so your team can search, navigate, and comment with precise timing.

How Transcription Works#

Eligible speech audio transcribes automatically after upload, subject to plan limits and provider availability. Video auto-transcription is off by default. An Admin can turn on Auto-transcribe video uploads in Settings → AI; with it off, start transcription manually from the version.

  • Background processing: Media is processed server-side, so you can keep working while it runs.
  • Progressive results: For large files, transcript text appears in chunks before the final pass finishes.
  • Word-level timestamps: Each word includes timing data for accurate seeking and comment ranges.
  • Platform-managed AI: Transcription runs on Cuevue's managed transcription service — no API key needed.

Speaker Labels#

When a transcript includes speaker information, Cuevue tags each paragraph with a color-coded speaker chip in its header, so a multi-person conversation is easy to follow.

  • Rename speakers: Click a chip to rename a speaker (for example, "Speaker A" → "Sarah"). The new name applies everywhere that speaker appears.
  • Reassign a paragraph: Open a chip's menu and choose Reassign section to another detected speaker if a paragraph was attributed to the wrong person.
  • Hide/show labels: Toggle speaker labels off if you'd rather read the transcript without them. This is a per-browser display preference.
  • Consistent colors: Each speaker keeps the same color throughout the transcript.

Not every transcript includes speaker data — it depends on how the media was processed. When speakers aren't detected, the transcript still has full word-level timing, search, commenting, and editing.

Use transcript search from the review header to find specific dialogue quickly.

  • Word or phrase search: Highlights all matches in the transcript.
  • Next/previous navigation: Jump through hits without losing playback context.
  • Auto-seek behavior: Moving through matches keeps you aligned with the right moment in audio.
  • Keyboard support: Cmd/Ctrl+F and Cmd/Ctrl+K open search.

Working with Transcript Content#

Transcript Commenting#

You can attach feedback to exact transcript ranges:

  1. Single-click transcript text to move playback and place the insertion caret.
  2. Double-click a word or drag-select text to create a comment selection.
  3. Add a comment from the selection toolbar/sidebar.
  4. Cuevue stores start/end timestamps and the selected quote.

Mentions and Collaboration#

Use @mentions in Comments to route specific moments to teammates.

AI-Assisted Field Generation#

Cue can fill AI Text, AI Select, and AI Multi-Select custom fields, and a project's Description, from the project's transcript and other field values. You set each field's prompt in Settings → Fields — see Custom Fields.

How It Works#

Generation runs from the project, not from the review player:

  • Quick generate: Click the generate button on an AI field — in a List View cell or the item detail panel — and Cue writes the value straight into the field. If the current value was typed by hand, or edited after Cue wrote it, Cue asks before replacing it.
  • Edit prompt…: Choose Edit prompt… from the generate button's menu (right-click the button, or ⌥-click it to open the prompt directly) to edit the prompt for this run, see whether a transcript is attached, and preview the result. Choose Accept to save it or Regenerate to try again. The same dialog lists previous generations for that field, and you can restore any of them.
  • Generate for all: Right-click an AI column's header (or use the Fields menu) to fill that field for every top-level project in the Space, up to 50 per run. Projects that already have a value are skipped unless you choose to overwrite them.
  • Fill automatically when empty: With this turned on, Cue fills the field while it's still empty and never overwrites a value. It runs when a project's transcript finishes (if Include transcript is also on — charged to the Space owner's AI credits), and once the last blank field the prompt uses gets filled in from any view in the Space (runs from that editor's browser and uses their plan's credits, like a click).

Cue always gets the Space name and Space description as context. A transcript isn't required for the manual paths: a field with Include transcript off works from the project's title, description, and other fields only.

Typical Outputs#

  • Episode descriptions and summaries.
  • Guest names and topics.
  • A category picked from your own option list (AI Select / AI Multi-Select only ever choose options you defined).

Who Can Generate and Limits#

  • Members, Admins, and Owners can generate, for fields they're allowed to edit. Guests can't, even when a role policy lets them edit custom fields.
  • Each person can run up to 60 generations per hour in a Space, and a daily limit is shared with Cue's other AI features.
  • Each new generation uses your plan's monthly AI credits. Re-running an identical request returns the saved result without using a credit. See Billing.

Edit Mode#

Transcript text correction changes the words. To cut media using a transcript selection or edit the timeline, see Editing, Clips & Recording.

Members and above can correct transcript text directly in the review view (Guests cannot, by default — see Guest capabilities).

  1. Enable Edit Mode.
  2. Click a word to edit it inline.
  3. Press Enter to save or Escape to cancel.
  4. Exit Edit Mode when done.

Common use cases:

  • Fixing names, brands, and technical terms.
  • Correcting misheard words.
  • Cleaning punctuation for readability.

Limits and Failures#

If transcription cannot run, Cuevue surfaces status in the transcript panel:

  • Missing audio
  • Temporary service/API issues
  • Monthly transcription-minute limits (with billing upgrade path)

You can retry from the same view once the issue is resolved.