How to cut several reels out of one batch recording session

Last updated 2026-09-22 · 8 min read

What is the short version?

Batch recording is the sensible way to make short video: set the camera up once, read four scripts straight through, stop. What comes out of it is a single file containing four videos that nobody has separated yet, which is the job Vizard Agent is doing here.

1. Give Vizard Agent the recording and the scripts.
2. Say how many separate videos are in it.
3. Ask for each as its own file with its own title.

What do you need before you start?

The recording, and whatever you were reading from while you made it. Vizard Agent can find the boundaries from the pauses alone if it has to, but the written scripts make them certain — and they give each finished reel a title without you having to write one.

• The recording. One file, several takes.
• The scripts. Pasted as text, in order.
• How many videos. So nothing is merged by mistake.
• The caption style. Applied to all of them.
• Whether retakes are in there. They usually are.

What do you type into Vizard Agent?

Paste the scripts in and say there are several separate videos inside the one file. The written text is what turns a guess about an unusually long pause into a certainty about exactly where the third reel ended and the fourth began.

Prompt

Variants worth knowing:

• "Several separate reels in this one recording." The shape of the file.
• "Here are the scripts." The boundaries, in text.
• "The same style on all of them." One treatment, applied per video.

What does Vizard Agent actually do?

Here is the Vizard Agent sequence on a single recording session that contained three separate reels. The boundaries get established before anything at all is styled, because every later decision in the job depends on exactly where they fall.

1. Checks the recording's length and parameters and prepares frames to review.
2. Builds an exact transcript with timings to find the edges of each reel.
3. Reviews real frames to see the composition, the background and where a caption can sit.
4. Recognises all the text with timings, including the pauses between the separate takes.
5. Checks the internal timings and the gaps between the three reels.
6. Looks at key frames at the boundaries and in the middle of each script.
7. Confirms where each take starts and ends and how much space there is around the face.
8. Prepares separate timings, titles and short visual inserts for the three reels.
9. Builds the title cards and checks the text and contrast against the footage.
10. Creates captions for each reel in a consistent style.
11. Checks the source voice level before the final encode so nothing is left quiet or peaked.
12. Renders three separate vertical reels with their titles, captions and inserts.
13. Reviews six frames from each and tests a higher-contrast caption style before re-rendering.

Step five is the step that makes this reliable. The gap between two takes is usually longer than any pause inside one, so the timings of the silences answer the question directly rather than by inference.

Step seven is what stops a reel starting on the end of the previous one. A take often begins with a breath, a glance down at the script and a half-second of settling, and a cut placed on the first word includes all of that unless the frames are checked.

Step thirteen is the ordinary loop. The first caption style is legible on most of the footage and not on the brightest part, so a higher-contrast version is tested on one reel before all three are rebuilt.

What does the result look like?

Three separate vertical files from Vizard Agent, each opening on its own title card, each starting on the first word of its script and ending on the last one, all captioned in the same style and levelled to the same loudness for posting.

When does this not work well?

Some sessions cannot be split cleanly at all. If you ran the scripts together without pausing between them, if you restarted mid-sentence, or if two takes overlap in subject, then the boundaries have to be judged by Vizard Agent rather than measured.

• No pauses between takes. Nothing to measure.
• Restarts mid-sentence. The retake has to be found first.
• Scripts that blur into each other. Two reels about the same point.
• No written scripts. Doable, but less certain.
• Background noise filling the gaps. The silences are not silent.

How do you fix a result that came back wrong?

Say which reel it is and which end of it is wrong. Because Vizard Agent holds the whole session as a list of boundaries rather than as separate exports, moving one edge does not disturb any of the other videos or their styling.

• "Reel two starts too early." That boundary is moved to the first word.
• "The third one has my retake in it." The retake is found and dropped.
• "They should be one video." Two boundaries are merged into one file.
• "The captions are hard to read." The style is changed across all of them.

How does Vizard Agent compare to doing it yourself?

By hand you scrub the timeline hunting for the moments you stopped talking, which is precisely the part of the process that makes people give up on batch recording. It is twenty minutes of finding boundaries before any actual editing has started.

Common questions

How many reels can one session hold? As many as you recorded. Vizard Agent finds each set of boundaries.

What if I did retakes? Say so and Vizard Agent drops the abandoned ones.

Do I need the scripts? No, but they make the boundaries certain and supply the titles.

Can each reel have a different title style? Yes, though consistency is usually the point of batch recording.

Will they all be the same length? No — each is as long as its own script takes.

Can it add different music per reel? Yes. Tell Vizard Agent which mood belongs to which.

What about horizontal versions? Vizard Agent can export both from the same boundaries.

Can it number the files? Yes, in the order they were recorded or the order you list.

What if one reel is much better? Vizard Agent can tell you which came out cleanest to post first.

Can it caption in another language? Yes, and the same style is applied across the set.

Does it level all of them together? Yes. Vizard Agent measures the source once and matches the exports.

Can I add a shared intro? Yes, and a shared outro, applied to each reel.

What if I recorded eight and only want three? Name the ones you want and Vizard Agent cuts only those.

How do I make the next session easier? Pause clearly for two seconds between takes; that is all Vizard Agent needs.