CapCut is a genuinely good editor and it is free for the parts most people need. If you are making story Shorts with reaction memes on the punchlines, it will do the job, and this is the actual workflow rather than the version tutorials show where the meme folder is already curated and the timing lines up first try.
Worth knowing before you start: Pro is now $19.99 a month, up from where it sat for years. Most of what is below works on the free tier, and it is worth knowing which parts do not before you commit to a subscription you may not need.
1. Get the script in and the voice out
Paste your script into a text track and use the built-in text-to-speech, or record and import audio. The built-in voices are limited and heavily used, so a lot of creators generate narration elsewhere and import a file. Either way, the narration audio is the spine of the timeline and everything else hangs off it.
Time: two to five minutes.
2. Auto-captions, then fix the auto-captions
Auto captions is one of CapCut's genuinely strong features. Run it on the narration track and you get word timings without doing anything.
Then you fix them, because synthetic speech confuses caption models in reliable ways: proper nouns, numbers, anything said quickly. Budget for a pass through the whole caption track, and style once and save the style as a template so you are not restyling every video for the rest of your channel's life.
Time: five to ten minutes the first time, three or four after you have a saved template.
3. Find the memes
This is the part nobody times, and it is where most of the forty minutes goes.
You have eight to twelve punchlines in a forty-second script. For each one you need a reaction image that actually fits the specific joke, not the general topic. So it is a folder if you have built one, or Google Images if you have not, and either way it is a linear search where you know the feeling you want and are trying to remember which meme carries it.
The failure mode is not finding nothing. It is finding something adjacent, using it because you have been looking for four minutes, and ending up with a video where the memes are vaguely on topic and never actually funny.
Time: fifteen to thirty minutes, and it does not get much faster with practice, because the bottleneck is recall rather than skill.
4. Place and time each one
Drag each image onto the timeline as an overlay, position it, scale it, set the in and out points against the narration waveform.
The timing tolerance here is tighter than people expect. A reaction meme wants to land on the beat where the line turns, not where the line begins, and being 300 milliseconds early reads as the joke being explained before it happens. You will nudge most of them at least once.
Keep the image clear of the bottom-left and top-right of the frame. That is where the platform's own interface sits, and a perfectly timed meme behind a share button does not land at all.
Time: ten to fifteen minutes.
5. Export
1080×1920, and check that captions and memes are inside the safe area on a real phone before you upload a batch of six.
The free tier exports without a CapCut watermark for standard edits, which is the main reason the free tier is enough for most story channels. Pro buys effects, stock, and cloud features that this format does not really use.
Time: two minutes plus the render.
Where the forty minutes actually went
Roughly five on voice, eight on captions, twenty on finding memes, twelve on placing them, two on export. The editing is not the expensive part. Searching your own memory for the right image, one punchline at a time, is.
That ratio is the whole argument for automating this step and not the others. Voice and captions are commodity features that every tool does competently. Meme selection is the part that takes the time and the part that decides whether the video is funny.
What CapCut cannot do here
There is no semantic search over a meme library in CapCut, and there is not meant to be. It is an editor: it will place and time anything you give it, precisely, and it has no opinion about which image belongs on which line.
That is a reasonable division of labor, and it is also why this workflow has a floor of about half an hour per video no matter how good you get at the editor itself. The bottleneck is not in the editor.
How Memecut does this
Memecut collapses steps 1 through 4 into a paste. Script in, and voice, word-level captions, and a real reaction meme matched to each segment come out, with the meme chosen by meaning rather than by tag or keyword and placed on the specific line rather than in the general vicinity.
The part worth comparing directly is step 3. A search over a library of real, already-circulating memes runs per segment, which is the same job you were doing from memory, minus the recall problem that made it take twenty minutes. Why that matching step is harder than it sounds has the mechanics.
If you would rather keep editing by hand, the parts of this article still stand on their own. The workflow above is what we did before building the tool.
What to remember
CapCut does all of this and the free tier is enough for it, so ignore the $19.99 unless you need effects this format does not use. Save a caption template, keep memes out of the two corners the interface occupies, and time the image to the turn of the line rather than its start. And know where the time goes: it is not the editing, it is the twenty minutes of trying to remember which meme fits. Watch what that looks like automated and decide whether it is worth doing by hand.