Skip to content

Edit Video by Transcript

Edit by Text turns the speech in a video into a transcript you can edit like a document. Delete a passage and that part of the video is cut out. Rewrite a passage and Glintia records the new line with an AI voice. It’s the quickest way to tighten an interview, a talking-head video or a screen recording with narration.

For trimming by hand on the timeline, see Editing Video Clips.

  • Uploaded videos only. Edit by Text works on a clip you uploaded or recorded, not on stock footage.
  • Keep the file to 25 MB or less. Transcription reads files up to 25 MB and stops with an “Audio file too large” message for bigger ones. Shorten or compress a long video before you upload it.
  • The clip needs speech. A clip with no words gives “No speech was found in this video.”
  • Where to find it. Click Edit by Text on the rail. It’s there in the Video Editor, Full Editor and AI Builder modes. You can also open AI Studio, choose the Video tab and click Edit by transcript.
  1. Select the video on the canvas. If the scene has no video yet, the panel shows Select video: click it and choose one of your uploads, or upload a new one. The clip is added to the scene and selected for you
  2. Open Edit by Text and click Transcribe video
  3. Wait while the panel shows “Creating an accurate timed transcript…”. The transcript appears as a list of passages, each with its start and end time

Transcription costs 0.25 AI credits per minute of audio, with a minimum of 0.25 credits.

A passage is roughly one sentence. A new passage starts after a full stop, question mark or exclamation mark, or once a passage reaches 18 words or 8 seconds. Passages are the smallest unit you can cut: you can’t remove a single word from the middle of one.

  1. Find the passage you want to remove
  2. Click its trash icon (Cut passage from video). The passage is crossed out and marked CUT. Click the icon again to restore it
  3. Repeat for every passage you want to remove
  4. Click Review cost & apply

Cutting doesn’t use any credits. When no passage in the transcript has been rewritten, Review cost & apply applies your cuts straight away, with no confirmation step.

Click into a passage and change the words. The passage is marked NEW VOICE. When you apply, Glintia:

  • generates the new line with an ElevenLabs v3 AI voice
  • mutes the original sound for that part of the clip and plays the new line in its place

Voice costs 0.5 credits for every 15 characters of rewritten text, at least 0.5 credits per passage. A 150-character line costs 5 credits.

Match the mouth to the new words (optional)

Section titled “Match the mouth to the new words (optional)”

Before you apply, tick Lip-sync changed speech to also redraw the speaker’s mouth so it matches the new line.

  • Glintia first checks a few frames of each rewritten passage for a visible speaker. The check costs 0.25 credits per rewritten passage.
  • Passages with no visible speaker are skipped and not charged for lip-sync.
  • Lip-sync costs 11.25 credits per second of each passage it runs on, so a 4-second line costs 45 credits. It’s the most expensive part of an edit, and it takes a few minutes.
  • If lip-sync fails on a passage, its credits are refunded and that passage keeps the new voice without the mouth change.
  1. Click Review cost & apply. With Lip-sync changed speech ticked, the panel first says “Checking changed shots for visible speakers…”
  2. The Confirm transcript edit box lists the ElevenLabs v3 voice cost, the Lip-sync cost with the number of eligible passages, and the Total
  3. Click Confirm & apply, or Cancel to go back
  4. Wait while the panel shows “Cutting video and generating changed voice passages…”. When it’s done, it says how many edits were applied

Credits are charged as each generation starts.

Applying an edit rebuilds the video from the original file:

  • The clip is split into scenes. Every stretch of video that’s left between your changes becomes its own scene, sized to fit it. The first stretch stays in the clip’s original scene. The others are added as new scenes at the end of the design.
  • Captions are added. An Edited Transcript subtitle track with the remaining words is added to the video. You can restyle it or remove it in the Subtitle panel. See Subtitles and Captions.
  • The clip is renamed “Edited by Text” followed by a summary, such as “2 cuts, 1 rewrite”.

The Ask AI box rewrites or shortens the transcript for you. It doesn’t change the video until you apply.

  1. To edit only part of the transcript, tick the box beside each passage you want changed, or click Select all for AI. With nothing ticked, the AI works on the whole transcript
  2. Type an instruction, for example “Shorten this, remove filler words, make the hook stronger”
  3. Click Edit selected text with AI (or Edit full transcript with AI)
  4. Read the suggestions. The AI may rewrite passages (marked NEW VOICE) or cut them (marked CUT). Change anything you don’t want
  5. Click Review cost & apply

Each Ask AI request costs 2 credits. Remember that any passage the AI rewrites will be re-voiced, and charged for voice, when you apply. If you only want cuts, ask for passages to be removed rather than reworded, or put reworded passages back to their original text before applying.

Every time you apply, Glintia saves a version. The Version history list shows each one with its summary and time, and marks the one in use as CURRENT.

  1. Select the edited clip and open Edit by Text
  2. In Version history, click Restore next to the version you want. It loads as a draft
  3. Review it, then click Review cost & apply to rebuild the video from the original file

The first entry, Original transcript, is always kept, so you can return to the uncut video at any time. Glintia keeps up to 20 versions per clip.

When you select an edited clip and open Edit Video, an Edited by Text box shows the number of edit versions and when it was last edited. Click Continue Edit by Text to pick up where you left off.

StepAI credits
Transcription0.25 per minute (minimum 0.25)
Cutting passagesFree
New voice for a rewritten passage0.5 per 15 characters (minimum 0.5)
Ask AI2 per request
Speaker check (only with lip-sync ticked)0.25 per rewritten passage
Lip-sync11.25 per second of each eligible passage

Edit by Text is available on every plan and draws on your monthly AI credits. See AI Studio and AI Credits.

  • “Select one uploaded video on the canvas first.”: select exactly one clip, and make sure it’s one you uploaded rather than stock footage.
  • “This slide has no video. Choose one from your uploads.”: click Select video to add one.
  • Transcription fails with “Audio file too large”: the file is over 25 MB. Upload a shorter or more compressed version.
  • “This edit would remove the entire video.”: you’ve cut every passage. Restore at least one.
  • “You have reached your AI credit limit for this month”: you’re out of AI credits. Cuts still need a transcript, so you’ll need credits to transcribe.
  • A rewritten line sounds like someone else: that’s expected, see the caution above. Cut the passage instead, or record a new take of your own and add it with Recording.