How do TikTok creators use Valmera?

Upload the raw take, then describe the video you want: "reframe to 9:16, cut the dead air and the umms, karaoke captions in the beast style, punch in when I get loud, music underneath." The agent performs the edit and you refine it with follow-up messages. Every reply is checked against what the system actually did, so the agent can never claim a cut, caption, or effect it didn't make.

Can it turn a horizontal video into a vertical TikTok?

Yes. Ask for 9:16 and the agent reframes the edit by crop, pad, or blurred-pad — crop to fill the tall frame, pad with bars, or pad with a blurred copy of your own footage behind it. Captions rescale to the new frame automatically, so nothing ends up half off-screen. Valmera reframes the frame you tell it to keep; it does not follow a moving subject around with automatic tracking, and it never upscales beyond your source resolution.

Does it make the word-by-word captions TikTok uses?

Yes — that style is karaoke captions, and it's one request. Words pop in one at a time as they're spoken, up to six words a line, with presets including karaoke and beast, 12 bundled fonts, any color plus a per-word highlight color, nine entrance animations, and per-word emphasis on the words that carry the punchline. Captions are burned into the exported video, so there's no separate SRT file.

Can it add punch-ins and speed ramps like a manual edit?

Yes. Ask for automatic punch-ins and the agent zooms in on the words you emphasized most, using audio-stress detection on your own delivery — zooms aim at any point of the frame, up to 2.5x. Speed is a separate request: any range from 0.25x to 4x, applied to whole sections, with voices kept at natural pitch and captions, music, and effects staying in sync. One honesty note on slow motion: frames are duplicated rather than AI-generated, so below about 0.6x the motion visibly steps.

Is there music I can use, or do I have to upload my own?

Both. Valmera ships a built-in library of 24 royalty-free CC0 tracks across eight moods — upbeat, chill, cinematic, corporate, dramatic, hip-hop, ambient, inspiring — and you can upload your own track (up to 50MB) instead. Either way the music ducks smoothly under speech via sidechain compression, and you can ask for a loudness master to −14 LUFS, the target the platforms expect.

Can Valmera post the video to TikTok for me?

No. Valmera exports a finished MP4 that you download and upload to TikTok yourself — there's no direct publishing, scheduling, or account connection. It also produces one deliverable per request rather than a batch: you describe one video, review it, export it, and ask again for the next one.

Does it work on my phone?

There's no native mobile app, but the studio runs in mobile browsers, so you can upload, chat with the agent, review the preview, and download the export from a phone. Uploads go up to 2GB or three hours, and the one-time analysis of each upload runs with visible progress — longer files take longer before the first edit.

Every account gets 20 free credits a day (they reset daily and don't accumulate) plus a one-time 150-credit welcome bonus, with no credit card. Plus is $20/month for 800 credits per billing cycle and Pro is $50/month for 2,400 — and subscribers keep the 20 daily credits on top. Edits charge credits in proportion to the AI work each turn actually takes, so a simple caption pass costs the least.

AI Video Editing for TikTok Creators

The edit is what separates a take from a post: vertical frame, word-by-word captions, a punch-in on the good line, a speed ramp through the boring bit, music that sits under your voice instead of on top of it. Valmera does all of that from plain-English requests — you upload the take and describe the post.

Edit Your Next Post

Upload the take, describe the edit, review the result. Free to try.

Start Editing Free →

See pricing →

What Short-Form Editing Actually Costs You

Reframing every horizontal take by hand

Timing captions word by word

Keyframing zooms on every good line

Speed ramps that knock the audio out of sync

Music that drowns your voice

Exports that sound quieter than everyone else

None of that is creative work. It's execution — and execution is exactly what an agent can take off your hands. You upload the footage once, the system builds a word-level transcript and detects shots and silences (the analysis runs with visible progress; longer files take longer), and from then on the edit happens in sentences.

The Vertical Frame, Handled

Most footage isn't born vertical. Ask for 9:16 and the agent reframes the whole edit using one of three methods: crop to fill the tall frame, pad with clean bars, or blurred-pad — a softened copy of your own footage filling the space above and below so the post reads as intentional rather than letterboxed. Captions rescale to the new frame with it, so text never lands half off-screen, and 1:1 and 4:5 are one request away when the same cut has to work on another feed.

Two honest limits worth knowing up front: reframing never upscales beyond your source resolution, and it doesn't motion-track a moving subject — you tell the agent what to keep in frame rather than relying on automatic subject following. The reframing docs cover how each method behaves.

Karaoke Captions, the Way Short-Form Wants Them

Captions carry short-form: most of your viewers meet the video muted. Valmera builds them from a word-accurate transcript and styles them from your description. Karaoke captions pop words in one at a time as they're spoken, up to six words on a line — the beast and karaoke presets are the loud, high-contrast short-form looks, with podcast and elegant for calmer edits, plus stacked, iridescent, chrome, editorial, fashion, luxe, and impact.

Everything about them takes direction: 12 bundled fonts, any color including a separate per-word highlight, sizes S to XL on a continuous scale, bottom, middle, or top placement, nine entrance animations, and per-word emphasis so the punchline word goes huge or glows while the rest stays calm. If the transcript mishears a name, fix the sentence in the editable transcript and the captions re-render from your correction. Captions are burned into the exported video — there is no SRT file to download. The captions docs list every option.

Punch-Ins, Speed, and Cuts That Feel Edited

The rhythm tricks that read as "professionally edited" are all requestable. Automatic punch-ins use audio-stress detection on your own delivery: the agent finds the words you hit hardest and zooms in on them, and you can aim a zoom at any point of the frame, up to 2.5x. Speed runs anywhere from 0.25x to 4x on whole sections — race through the setup, sit at normal speed on the payoff — with voices held at natural pitch and captions, music, and effects staying in sync. Slow motion duplicates frames rather than synthesizing new ones, so below roughly 0.6x the motion visibly steps; keep the deep slow-mo short.

For the cuts themselves there are seven transition styles — dip to black, dip to white, whip left, whip right, zoom punch, glitch, and flash — and the style you choose applies at every cut in the edit. Being straight about it: there are no crossfade dissolves, and you can't give each individual cut a different style. You can also ask for a sound-design pass that drops whooshes on the cuts, an impact on the strongest word, and a riser into the energy peak, or ask the agent to snap the cuts to the beat of your music.

Sound That Doesn't Get You Skipped

Music comes from a built-in library of 24 royalty-free CC0 tracks across eight moods — upbeat, chill, cinematic, corporate, dramatic, hip-hop, ambient, inspiring — or from a track you upload yourself (up to 50MB). Either way the agent ducks it under your voice with sidechain compression, the way a mix engineer would, instead of parking it at a flat low volume. Ask for a loudness master and the export lands at −14 LUFS and −1.5 dBTP, the targets the platforms normalize to, so your post doesn't arrive quieter than the one before it. The audio and music docs go deeper on gain, fades, and voiceover.

Requests TikTok Creators Actually Send

Horizontal take → vertical post

Reframe this to 9:16 and keep me centered in frame. Cut the silences and the umms, add karaoke captions in the beast preset with a yellow highlight word, and punch in whenever I get loud.

Pacing pass

Speed the setup at the start up to 1.4x, keep my punchline at normal speed, and drop the last three seconds to 0.75x. Put a whip-left transition on every cut.

Sound

Put an upbeat track from the library underneath, duck it under my voice, and master the whole thing to −14 LUFS so it isn't quieter than everything else on the For You page.

Send them one at a time and review between each — that iteration is the workflow, not a workaround. The Free plan's 20 daily credits are enough to run all three on a real take; posting daily fits the Plus plan at $20/month for 800 credits per billing cycle, with the 20 daily credits still arriving on top.

What Valmera Won't Do for You

Three limits, stated plainly, because finding them out later is worse. One deliverable per request: each turn produces one program, one preview, and one export — describe a video, review it, export it, then ask for the next. There is no "give me ten clips from this stream" batch mode. No direct publishing: you download the finished MP4 and upload it to TikTok yourself; Valmera doesn't connect to your account, post, or schedule. No native mobile app: the studio runs in mobile browsers, which covers uploading, chatting, previewing, and downloading from a phone, but it isn't an app in the App Store. If bulk clip batches are your entire workflow, the Opus Clip comparison is the honest read; if you want the tall-frame edit done to your direction, that's this.