How to make a two part reel that ends part one on a cliffhanger

Last updated 2026-09-04 · 7 min read

What is the short version?

A two-part reel is not one video cut in half. It is one story with a deliberate hole in the middle — part one has to stop somewhere that leaves a question open, and part two has to answer it without spending its first eight seconds explaining what part one already said.

1. Go to Vizard Agent and upload the recording.
2. Say you want it as a two-part reel.
3. Say what the story is, so it knows what the turn is.

What do you need before you start?

One continuous telling of the thing. Vizard Agent needs the whole arc in the footage to find the point where it bends — you cannot split a story into two parts if the recording only ever contains the first half of it.

• The footage. One telling, all the way through.
• The story. In a sentence, so it knows the shape.
• The platform. It sets the frame and the length.
• A caption style. Or let it audition some.
• Any hard length. Sixty seconds each, or shorter.

What do you type into Vizard Agent?

Say "two-part reel", not "two reels". The phrasing carries the whole instruction: two reels means two independent videos, and a two-part reel means one story deliberately broken at a point that makes people come back for the rest of it.

Prompt

Variants worth knowing:

• "Part 1 ends on the problem." The turn, named for it.
• "Same look on both." They have to read as a pair.
• "Label them." A badge in the corner saying which is which.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real two-part reel cut from a personal story told to camera. Step eight is the decision the whole thing rests on, and it is made from the transcript rather than from the picture.

1. Transcribes the footage and extracts word-level timings.
2. Reads every transcript segment to see the shape of the telling.
3. Checks the second clip's details and samples frames from both.
4. Measures the audio levels of the source.
5. Builds a full word-level transcript in readable form.
6. Checks the speech gaps to find the natural pauses.
7. Builds a contact sheet across the footage and looks at it.
8. Analyses the cut points for part one against the transcript.
9. Builds rough cuts of both parts with jump cuts at several framings.
10. Checks their durations and uploads them for validation.
11. Verifies the transcript and the continuity across both cuts.
12. Renders refined jump cuts for both reels.
13. Searches and measures a music bed.
14. Re-transcribes the clean cuts for final word timings.
15. Generates animated captions for each part separately.
16. Auditions several caption presets on real frames.
17. Builds the overlay badges and rebuilds the ones that read poorly.
18. Renders both parts, builds contact sheets, and reviews them.
19. Measures the loudness of both so they match on playback.

Step eleven is the one people skip and regret. It is easy to cut part one at a good moment and then discover part two opens on the second half of a sentence, or repeats a line that already went out. Vizard Agent re-reads both transcripts against each other before committing.

Step fourteen is quietly important too. The captions are timed against the cut video, not the original — timings taken from the source and shifted by hand are where two-part reels usually go out of sync.

What does the result look like?

Two vertical reels that read as one thing. The same caption treatment, the same framing rhythm, badges marking which is which, and levels matched so part two does not arrive louder than part one when someone plays them back to back.

Part one stops where the story turns rather than where a timer ran out. That is the whole difference between a series and a video someone chopped up.

When does this not work well?

The method assumes there is a turn to find. Plenty of good footage has no bend in it — a list, a tour, a demonstration — and splitting that produces two halves rather than two parts, which viewers notice immediately.

• Stories with no turn. A list does not have a cliffhanger in it.
• Very short recordings. Two thirty-second halves need real material.
• Several topics in one take. The split becomes arbitrary.
• Footage that repeats itself. Part two feels like a rerun.
• A punchline in the first ten seconds. Nothing left to hold back.

How do you fix a result that came back wrong?

Tell it where the split should have been. Vizard Agent keeps the word timings, the caption files and the rendered assets for both parts, so moving the break rebuilds the two cuts without re-transcribing anything or restyling the captions from scratch.

• "Split it later." Both parts re-cut from the same timings.
• "Part two repeats a line." The overlap is trimmed at the join.
• "They look different." One caption style applied across both.
• "Part one gives it away." The reveal moved into part two.

How does Vizard Agent compare to doing it yourself?

By hand the hard part is not the cutting, it is reading a transcript looking for the moment the story turns — and then keeping two exports consistent when you inevitably restyle the captions on one of them and forget the other.

Common questions

Can it do three parts? Yes. Vizard Agent splits at as many turns as the story actually has.

Does part two need its own hook? It gets one. Vizard Agent opens it on a line that stands alone.

Will it label them? Ask and it will. Vizard Agent builds a corner badge for each part.

Can I choose the split point? Yes. Give a timecode or a line and Vizard Agent cuts there.

What if my story has no turn? Vizard Agent will say so. Two parts is not always the right shape.

Do both parts get the same captions? Yes. Vizard Agent styles once and applies it across both.

Can it make part one shorter? Yes. A shorter part one and a longer part two is a common shape.

Will the audio levels match? Yes. Vizard Agent measures the loudness of both renders.

Can it use different footage for each part? Yes. Upload both and Vizard Agent treats them as one story.

Does it re-transcribe after cutting? Yes, for the captions. That is what keeps them in sync.

Can it post them as a series? Vizard Agent produces the files and can schedule them. Posting stays yours.

What if part two is weaker? Say so and Vizard Agent rebalances the split.

Can it add a "part 2 in comments" end card? Yes. Name the wording and Vizard Agent builds it.

Why not just cut it in half? Because the middle of a recording is almost never the middle of a story, and viewers can tell the difference within a second.