Character switch

Your take. Your timing. Their face.

Film yourself delivering the lines, for anything from two seconds to ten minutes. heyCreatr tightens the take, finds the short videos hidden in it, films it again from new camera angles, and keeps your performance and, if you like, your voice, while any character from your library plays the part. The way to be on camera without being on camera.

Record

Up to ten minutes and 1 GB, face visible. MP4, MOV or WebM from your phone is fine, with a progress bar you can cancel.

Clean up

Long pauses and filler words are found and cut for you. Keep or undo any cut, then stitch a tighter take.

Cut and cast

The take is cut at the pauses between your sentences. One character for the whole take, a different one per piece, or your own footage kept.

Make the video

Every piece becomes a scene in the performance editor, with your voice track, captions, B-roll, overlays and text on one timeline.

New

From one long take to a week of posts

Film once, talk for ten minutes, and let heyCreatr do the cutting you would otherwise spend an evening on.

✂

Clean up

Pauses and the ums and ers are found in the take, in any language, and shown as cuts on a waveform. Choose natural, tight or very tight, keep or undo each cut, listen to the result, then stitch it into a new take. The original is never touched.

✦

Reels

Choose 15, 30 or 60 seconds and heyCreatr suggests the stretches that stand on their own: a hook at the start, one point, a payoff. Play each one, then make it a reel of its own, cut from your take, in 9:16 with captions on.

▶

Auto edit

One click for a finished cut: it opens on your strongest moments with a title, flashes, and plays the take from the start without pauses or fillers, with bold captions, camera rhythm, stock B-roll and music if you want it.

Reels and auto edits are new videos under the one they came from. The original is never touched, and everything they add stays editable. Join finished videos into one film in the Cutting Room.

New

One phone, more camera angles

Film a static shot and an AI camera films it again from another angle: the same people, the same room, the same words, a new camera. A three-quarter view from the side, close-ups on whoever speaks, over the shoulder, a wide shot, or your own words.

  • Your voice stays yours. The new angle plays under your own recording, and in our tests the words landed within a few milliseconds of the original.
  • People stay themselves. heyCreatr finds everyone in the take, and each one gets a character card with a master sheet, so their face holds from a side the camera never saw.
  • Priced before it runs. Adding an AI camera costs nothing. A scene is filmed when you switch it to that camera, once, and kept.
  • Cut between them like real cameras. Pick the camera per scene, or cut between the main take and the AI camera in one go.
Length2 seconds to 10 minutes
UploadUp to 1 GB, MP4, MOV, WebM
Pieces5 to 30 s, auto or by hand
CamerasUp to 3 phones and 3 AI cameras
Reels15, 30 or 60 s, up to 8 at a time
VoiceYours, another voice, or none
Example cost6-second swap, $0.50

Long takes

Cut where you pause, automatically or by hand

  • Auto: pick pieces of 5, 10, 15, 20 or 30 seconds. Every cut lands in a pause, never inside a word.
  • Manual: a waveform and word timeline. Press C to cut at the playhead, drag a cut to move it, nudge it a frame at a time, or type its time to the millisecond.
  • Snap to pauses keeps cuts between words, and zoom goes from the whole take down to about a second across the screen.
  • Every saved set of cuts is a new version, so videos made from earlier cuts keep working.

Real cameras too

Two or three phones, lined up on the sound

Film the same take on two or three phones, clap once, and heyCreatr lines them up on the sound. Real cameras cost nothing to switch between, and they sit next to your AI cameras in the same camera list.

Everything the performance editor gives you

⇆

A character per piece

Click pieces to recast them, or keep your own footage in any of them. The panel counts and prices only what is swapped, before anything runs.

✦

Test one piece first

Swap a single piece to compare models at the price of that piece alone, then run the whole take.

☺

The look holds

Each piece starts from the last frame of the piece before, so the character does not drift over a long take.

▣

A camera per scene

Main, a second phone or an AI camera: pick one per scene on the scene card, or cut between them all in one go, with the price shown first.

▶

Camera moves

Push in, pull out, zoom, pan, tilt, punch-in, slow drift or handheld, with an amount. Free, instant, and the lips stay in sync.

☖

A new room

Draw one set from your own frame: an office, a studio, a café, or your own picture. Put the whole video in it, or just one scene. About $0.08 to draw.

♫

Your voice or another

Keep your own voice, or give the whole video another one: the same message retold in fresh words in any voice, or your exact words and timing in another voice through your ElevenLabs account. Listen to any voice before you pick it.

∿

One voice track, captioned

Keep your voice and the whole take runs as one track under every scene, with captions on every spoken word, so B-roll can cut away while you keep talking. The music ducks while you speak.

✎

Recast any scene later

Change who plays one scene, split a scene in two, or put your original footage back, with the price shown first.

B-roll

Cut away without losing your voice

Every sentence in the transcript has B-roll here. It adds a cutaway of up to five seconds over that sentence while you carry on talking underneath.

  • Drawn by AI: leave the prompt empty and it pictures the sentence itself, or describe the shot you want. Animate a still into a short clip, or let it move with Ken Burns.
  • Real stock footage: clips and photos from Pexels and Pixabay, free for commercial use, with your own free keys. Find matches writes the search terms from your sentence, in any language, and credits the filmmaker.
  • Fill a whole video: add B-roll to every sentence, every second or every third in one go, then only swap what you don't like.
  • Speed per clip: play a clip from a quarter to eight times its speed, or fit the whole clip into its slot, so a ten-second shot plays in two.
  • Split screen: put the B-roll beside you instead of over you, left, right, top or bottom, taking 30 to 70 percent of the frame.
  • Your own: upload an image or a clip.

Overlays and text

Badges, product shots and titles that move

  • Overlays: a badge, an icon or a product shot, drawn by AI on a transparent background.
  • Placed to the percent: across, down, width and opacity, with its own start and end on the timeline.
  • Motion: an entrance, an exit and a loop, from fade, slides, pop, float, pulse and spin.
  • Text: put any words on screen in modern type, from editorial serifs to bold condensed, with the same motion.
  • What you see is what exports: the preview and the MP4 use the same motion.

Straight about the limits

Ten minutes and 1 GB per take. The swap models work on 30 seconds at a time (10 for the video editors), so a ten-minute take is about twenty swaps, and the panel shows the count and the cost first. Reels and auto edit pick moments from what you say; they are suggestions, so watch them before you post. An AI camera films scenes of four seconds and more, from a take filmed on a steady phone, and a face the take only ever shows from one side is partly completed by the model, so check its master sheet. Face-only swaps arrive when the model catalogue has one. And you confirm you have the right to use the footage: the people on camera are you, or have agreed.

Get in among the first

heyCreatr isn't open for signups yet. Join the waitlist and we write the moment it opens.