How to Record and Trim AI Videos With Tella
Tella is a browser-based recorder for screen, camera, or both, built for people who need to explain something on video without learning an editor. The recording itself is ordinary. The useful part is what happens after: it strips silence, tags the spots where you fumbled, and lets you delete a sentence from the middle without re-recording.
A three-minute product demo that used to take an afternoon of shooting takes about thirty minutes, most of it being the editing pass where you cut the tangents.
How does the recording work?
You open tella.tv, choose screen plus camera, and hit record in a browser tab — no install needed, though there is a desktop app for higher frame rates. The screen and your camera are recorded as separate layers, so you can resize your face into a corner or hide it entirely later.
Before recording, set the canvas size. Most Tella videos are 16:9; if you are making a vertical clip for social, set 1080×1920 first, because changing aspect ratio later crops your screen content.
What does the AI editing pass actually remove?
| Feature | What it does | How good it is |
|---|---|---|
| Silence removal | Cuts dead air over ~0.5s | Reliable |
| Filler word tags | Flags “um”, “so”, “like” | About 80% clean, misses some |
| Error detection | Marks re-takes of the same sentence | Useful, needs a check |
| Auto captions | Generates subtitles from audio | Accurate on clear mic audio |
Treat the fillers as suggestions. Blindly accepting every cut produces a jumpy rhythm — speech with all ums removed sounds oddly fast. Cut about half of them and the delivery sounds natural.
How do you cut a paragraph from the middle of a recording
Selecting text in the transcript deletes the corresponding video section. That is the workflow worth learning: read the transcript, delete the sentence that goes off-topic, watch the joint, and if the jump is harsh, add a small zoom-to-cursor transition.
Three editing rules that hold up:
- Cut on the end of a sentence, never mid-word.
- After every cut, watch two seconds either side.
- Re-record a whole section rather than patching three separate mistakes into it.
Is Tella free?
There is a free plan with a watermark and a cap of around five videos. Paid tiers start near $19 per month and remove the watermark, add unlimited recordings, and raise the video length cap. For a team of five, the cost of three plans is roughly what one hour of an editor’s time costs in most markets.
When a screen recorder is the wrong choice
If the video needs tight animation, motion graphics, or a scripted narrative, this is the wrong category of tool. Tella records what you show it — it will not make an unpolished demo look designed. It is also the wrong tool for recording a live meeting with many speakers; that is what a meeting recorder does.
Where it wins: async product walkthroughs, bug reports recorded instead of typed, onboarding videos for internal tools, and sales demos that need to be personalised per prospect. Teams that send five recorded explanations a day get their money back in a week.
For turning the finished recording into text-based clips with stock footage, Pictory Video covers the generation side. For narration-only explainers where no screen needs recording, generated b-roll from MiniMax Hailuo is the cheaper route. And if you script the narration first, dictating it with Wispr Flow Dictation keeps the delivery starting clean.
Quick Summary
- Tella records screen plus camera in a browser, no install required.
- Automatic silence removal is reliable; filler-word cuts should be accepted selectively.
- Delete a whole sentence from the transcript to remove a tangent without re-recording.
- Free plan caps at about five videos and adds a watermark; paid starts near $19/month.
- Best for async walkthroughs and bug reports, not scripted or animated content.
