cutipAI Logo
BlogStart your project

Templates

Ready-made jobs you can run without opening a project — camera moves, video edits, talking images, speech cleanup, virtual try-on and product holding.

A template is a job that is already set up: you bring an image and a few choices, and it does the rest. Templates run outside any project.

Finding a template

Templates are grouped by what they make: Image, Video and Audio. The switch at the top of the page shows how many templates each category holds. All lists every template under a heading for its category; choosing a category shows only those templates. The page address remembers your choice, so you can bookmark or share a category, and Back returns to the one you had before. Closing a template returns you to the category you opened it from.

Edit video

  1. Open Edit video under Video to change a clip with written instructions. Upload a video or choose one from Sandbox or a project, then describe what should change. The prepared MP4 must be under 15 seconds and under 50 MB; longer footage opens a trimmer so you can choose a part. The result is a 720p video.

    Hover or focus Generate to review the estimated cost before starting. After generation, compare the source and edited result with shared playback controls and download the result. The source video remains unchanged.

Describe the change you want in Edit video, then choose a source clip.
Describe the change you want in Edit video, then choose a source clip.

Upscale video

  1. Open Upscale video under Video and choose a clip up to 20 seconds and 50 MB. Precise is selected initially and keeps enhancement faithful to the source. Creative can invent new detail and accepts optional enhancement instructions.

    Choose a Scale factor from 1.5× to 3×. The source aspect ratio is preserved, and the form shows the predicted output dimensions. Verified estimates are available for 1920 × 1080, 2560 × 1440 and 3840 × 2160 output, including portrait equivalents. Other output sizes cannot start until their pricing is verified; adjust the scale or source instead. Trimming or converting for this template preserves source dimensions. Current outputs use 24 frames per second.

    Review the estimate on Generate, then compare and download the result. Both new templates save their source and settings before submission. Open in Templates on a saved result restores its recipe; generating again creates a new paid result. Submitted work remains recoverable in History after you leave the template.

Upscale video starts in Precise mode with a 2× scale factor.
Upscale video starts in Precise mode with a 2× scale factor.

Remove background

  1. Open Templates → Image → Remove background and upload one JPEG, PNG or WebP, or choose an image from your Library or a project's generated media. Images must be below 20 MB and at most 32 megapixels. The cover is illustrative artwork, not a processed example.

    1. Choose Automatic to detect the subject, Prompt to describe what to keep (for example, “keep only the person and their bag”), or Remove a color for a solid-color background.
    2. Automatic and prompted removal show their estimated cost before Remove background. Prompted removal has a different price. Color removal runs locally with no AI charge: use Pick color from image, Color tolerance, and Edge softness. Matching colors inside the subject can also disappear.
    3. Review the checkerboard preview with Zoom and Before / after. Use Erase and Restore to correct transparency, with brush size and softness controls. Restore recovers original pixels, retaining any transparency in the original. Undo and Redo work on corrections, key settings and the background fill; they do not reverse a generation charge. Changing the key settings replaces brush corrections, so adjust the key before brushing.
    4. Keep Transparent or choose a solid Background, custom color or hex value. Download PNG exports at the original dimensions. Save to Library keeps a separate finished variation and its transparency mask, so you can reopen it through Open in Templates.

    The original paid cutout stays in your Library. Check original job retrieves an interrupted request without submitting another paid job. Start a new cutout unlocks the source and starts a separate request. Unsaved manual edits remain in the current session; download or save them before leaving. Brush history keeps up to 20 steps within a memory budget, so large images retain fewer steps. Removing a color, brush corrections and background fills require no additional AI job.

Relight

  1. Relight's split-light portrait cover illustrates the lighting change; it is cover artwork rather than a processed example.

    Relight changes a video's lighting using a sphere image that shows the desired light, colour and shadows. The lighting preview is approximate; generate to see it applied to your video.

    1. Open Templates → Relight and upload a video, choose a saved video, or provide a video URL. The whole video must be between approximately 2.34 and 15 seconds, up to 50 MB. Trim longer footage first; this studio does not split videos or create paid batches.
    2. Choose Lighting studio or Upload sphere image. Upload accepts JPEG, PNG or WebP sphere references up to 20 MB. In the studio, choose a preset, select a light and drag the sphere view using Horizontal orbit, Vertical orbit, or Distance. Numeric controls set exact positions. The degree buttons set azimuth in 45° steps.
    3. Use one to four lights. Adjust each light's Colour, Intensity, and Softness, plus Ambient fill for the scene. The sphere image, without the guide markers, is sent as the lighting reference.
    4. Choose 768p (default), 480p, 1080p, or 2K, and an aspect ratio. Match source is the default. Leave Seed blank for a random variation. Hover or focus Relight video to inspect the estimated cost before generating.
    5. Generate and use Play comparison to review both videos together. Original sound plays; result sound is muted to prevent doubled audio. Download the result when ready.

    Generation converts the source to 24 fps and can shorten its end to a supported frame count. The studio shows the prepared duration before generation. Final charges follow actual generation usage. Start a new relight starts a separate paid request; Check original job retrieves the same interrupted request. Saved results can reopen their recipe through Open in Templates.

    In a project, right-click a video in Timeline or Assets and choose Relight…. A timeline clip's visible playback must fit the duration limit; its source file can be longer. Speed changes, reverse playback, and freeze frames are supported. Relighting processes playback only, keeping effects, framing, keyframes and transitions editable.

    Apply relight replaces only the targeted timeline clip in one undoable edit. Its start and end stay fixed; the studio shows the speed adjustment used to fit the result and previews that timing. Existing timeline audio stays unchanged. If the source, clip or track lock changed, Apply refuses and keeps the result in Assets. Undo restores the original clip and retains the generated video. From Assets, Keep in Assets retains a new video without replacing existing assets or timeline uses. Closing the studio never applies a result.

Recast your video

Recast your video replaces people using your photos while aiming to retain their performance, camera movement, cuts and source sound. The template tile previews the original and recast performance side by side; reduced-motion users see a still frame.

  1. Open Templates → Recast your video, then upload a video or choose a saved generated video.
  2. Set In and Out to select 5–30 seconds. You can scrub the preview and use Set In here and Set Out here, or enter seconds directly. Longer source videos are allowed up to 50 MB.
  3. Review Detected shots. Every shot must be at most 15 seconds; an overlong detected shot blocks generation. Shorten the range to resolve it. Detection estimates cuts and can miss an edit; the generation service validates the footage.
  4. Choose 1–4 JPEG, PNG or WebP replacement photos, up to 20 MB each. Use Move earlier and Move later to order them. The default mapping replaces the main people from left to right. For a moving group, use Replacement instructions to identify who becomes whom and what to keep.
  5. Choose 1080p (the default) or 768p. Leave Seed blank for a random variation, or enter a nonnegative whole number. Hover or focus Recast video to inspect its estimated cost.
  6. Generate, then use Play comparison to view the source range and result together. Choose Source sound or Result sound; only one plays at a time. Download the result when ready.

Price depends on selected duration and resolution; 1080p has a higher rate. Extra replacement photos add no surcharge. The final charge follows actual generation usage.

In the editor, open a video clip's menu and choose Recast people…. This currently supports normal forward playback at speed 1. The result is added to Assets before any edit changes. Apply selected range replaces only the selected interval, splitting the clip when necessary. Surrounding footage, effects, keyframe motion and existing timeline audio stay in place. Apply is one undoable edit; undo keeps the generated result in Assets. If the source or edit changed, Apply refuses and retains the result for manual use.

Closing the dialog keeps its job available from the clip menu. After an interruption, use Check original job to retrieve the same request. Start a new recast creates a separate generation with a separate charge.

Multi-angle

Multi-angle turns one photo into a moving shot. You set where the camera starts, where it ends, and how it travels between them.

  1. Click Templates in the menu on the left of your account pages, choose Video at the top if you want to narrow the list, then choose Multi-angle.

    The Templates page.
    The Templates page.
  2. Choose the image to start from:

    • Upload — an image from your computer.
    • Sandbox — an image you made in the Sandbox.
    • From a project — choose a Project, then one of the images generated in it, including first, middle and last storyboard frames.

    To start again, click Choose a different source.

    Multi-angle: the source and prompt on the left, an example on the right.
    Multi-angle: the source and prompt on the left, an example on the right.
  3. In Prompt, write what should stay true across the move — the subject, the place, the light.

  4. Set the camera's path. Each pose has an Azimuth (around the subject), an Elevation (above or below it) and a Distance. Click Add pose for another stop on the way, and Play to preview the move.

  5. Choose a Duration and a Resolution.

  6. Click Generate. The result plays beside the form when it is ready.

Only Duration and Resolution change the price. Adding, moving or removing poses never does. Resolution is the expensive choice: 1080p costs about twice as much as 720p.

Prompt rewriting can polish your prompt before it runs. Leave it on Balanced, choose Maximum quality, or choose Off — use exactly what I wrote. When it is on, the result shows the version that actually ran, under Ran as.

Restyle a video

Restyle a video gives an existing clip a new look, frame by frame, while keeping its motion and framing.

  1. Choose the clip to restyle:

    • Upload — a video from your computer.
    • Sandbox — a video you made in the Sandbox.
    • From a project — choose a Project, then one of the videos generated in it. Videos you edited in that project appear here too.

    The clip can be any length — a longer clip is trimmed. MP4, MOV and WebM clips are converted for you when needed.

  2. Pick a look — Anime, Claymation, Film noir, Watercolour, Oil painting, Pixel art, Snowy winter or Night-time — or write your own under What should change?. Editing a look's words moves the choice to Custom.

    Choose Match a picture to use a picture's visual style instead. Add the Picture to match from Upload, Sandbox or a project, then optionally describe Anything else to change?

    Restyle a video: the look chips and prompt on the left, a before-and-after example on the right.
    Restyle a video: the look chips and prompt on the left, a before-and-after example on the right.
  3. Hover or focus Generate to see the Estimated cost, then click it. The estimate includes the input and output lengths for looks that bill both. Final charges follow actual generation usage. Some looks take a few minutes.

Anime, Claymation, Watercolour and Match a picture accept clips from 3 to 30 seconds. The prepared picture must be 700 to 4,553 pixels on each side, with a shape between 2:5 and 5:2, and run at 24 to 60 frames per second. Small clips are enlarged during preparation. If your chosen part is shorter, choose another look or use Change the part to select a longer part.

Longer clips

Longer clips open a trimmer. Anime, Claymation, Watercolour and Match a picture use up to 30 seconds; other looks and Custom need a part under 15 seconds, with up to 14.5 seconds available in the trimmer. Drag the highlighted part or its handles, use Play the part to check it, then Use this part. Switching looks updates the limit without converting the clip again. Only that part is converted, priced and sent — so a shorter part costs less. Change the part takes you back to the trimmer.

What is sent must be under 50 MB once converted. A part that is still larger is not sent: the trimmer stays open and asks you to Choose a shorter part.

Restyle a video with a longer clip: the trimmer under the preview, with the chosen part highlighted and its start, end and length below.
Restyle a video with a longer clip: the trimmer under the preview, with the chosen part highlighted and its start, end and length below.

The result is a new video, and your original clip is kept. Click Try another look to run the same clip again with a different look, without choosing it again. The result also lands in your Sandbox list, alongside everything else you have made there.

Reframe a video

Turn one clip into other shapes — vertical for Stories and Reels, square, a feed post, or cinema — without opening a project. Your whole picture is kept and the new edges are filled in.

Reframe a video: your clip, resolution and shapes on the left, a before-and-after example on the right.
Reframe a video: your clip, resolution and shapes on the left, a before-and-after example on the right.
  1. Open Templates and choose Reframe a video. Add a clip under Upload, Sandbox or From a project. MOV and WebM clips are converted for you. A clip longer than 60 seconds opens a trimmer so you can choose the part to use.

  2. Choose a Resolution: 1080p or 720p, or 540p for a clip of up to 10 seconds, a lower-cost choice that also offers 21:9 — Cinema, 4:3 and 3:4.

  3. Tick every shape you want. Each one is named for where it is used, such as 9:16 — Stories, Reels, Shorts. Your clip's own shape is not offered.

    A clip chosen: pick a resolution and tick every shape you want.
    A clip chosen: pick a resolution and tick every shape you want.
  4. Click Generate. One video is made per shape, and each appears beside your original as soon as it is ready, with a Download button.

Before you add a clip, the result area plays an Example: a film reframed to 9:16, 1:1 or 21:9, beside its original. Click Use these settings to load that example's resolution and shape — then choose your own clip.

Each shape is its own generation: if one cannot be made, the others still arrive, and you are charged only for the ones that land. Every result is also saved in the Sandbox history. A shape that takes far longer than it should — about ten minutes plus a quarter of a minute for each second of your clip — is stopped, and you are not charged for it; try again, or choose a lower resolution.

Looping video

Choose Looping video to turn one JPEG, PNG or WebP image into a seamless loop. Upload an image, use a public image URL, choose a saved Sandbox image, or select a generated project image.

Choose Subtle motion for gentle ambient movement or 360° spin for a full revolution. The camera stays locked in both modes. Motion instructions are optional.

Set Duration from 5–15 whole seconds and Resolution to 480p, 768p or 1080p. The defaults are five seconds, 1080p and subtle motion. Include audio is off by default; enable it to keep generated audio. Duration and resolution determine the estimated cost, plus the usual fee.

Hover, focus or press and hold Generate to see its estimate. Generation requires a selected image, valid settings and an available quote. Inputs stay locked while the job runs; Cancel stops queued or running work. Completed outputs are kept while saving. If the image upload or generation is refused, the error message explains the reason so you can correct it before retrying.

Play the completed video on repeat or choose Download video. The source image, settings and result are saved in your Sandbox library. Interrupted jobs use the existing request to check for a result rather than starting another paid generation.

Talking image

Choose Talking image to animate a face image with your speech audio. This creates a new video from a still; the project's Improve lipsync control continues to work on existing video.

Its tile previews a real generated result. Open the template to play the full example with speech, or choose Use these settings to load its resolution, transcription choice and seed. Your selected image and audio stay in the form. If you prefer reduced motion, the tile shows a still frame instead.

Choose a Face image from an upload, a public image URL, the Sandbox or a project's generated images. Use a clear, unobstructed face. The image ratio must be between 0.4 and 2.5; the output uses the nearest supported ratio.

For Speech audio, upload a recording, choose saved Sandbox audio or select generated dialogue from one project. Listen to the audio preview and check its displayed length. Recordings must be between 5 seconds and 15 minutes, and no larger than 50 MB. Short recordings are not padded and longer recordings are not trimmed.

Choose Resolution: 480p, 768p, 1080p or 2K. The default is 768p. Under Additional settings, you can turn on Transcription guidance to use the transcribed speech to guide mouth movements, or enter a Seed. Leaving the seed blank chooses a random one.

Hover or press and hold Generate to see its Estimated cost. Duration and resolution change the estimate; recordings longer than 15 seconds have a duration surcharge. Generation is available once the inputs and estimate are ready. The server verifies the audio length before reserving funds; the final charge follows actual billed usage plus the usual fee. Usage billing can remain pending briefly after the video arrives.

Click Generate, then play the completed video with sound or choose Download video. Cancel stops a queued or running request; completed outputs are retained while saving.

Clean audio

  1. Choose Clean audio and upload an audio or video recording, or choose saved media. Recordings can be up to 30 minutes and 512 MB. The result is 48 kHz mono audio. This tool is for speech: background music may be removed too. Video inputs produce audio only.

    Under Advanced, adjust Strength or turn Normalize loudness off. Hover, focus or press and hold Clean audio to see the estimated price. Billing rounds the input to whole seconds, then charges whole minutes with a one-minute minimum, plus the usual fee.

    Switch between Original and Cleaned to compare the result, then download it. Results and settings are saved in your Sandbox library. Open in Templates restores a saved cleanup; Check original job retrieves an interrupted request without submitting another paid job.

Where the result goes

Every template result is kept in your Sandbox, alongside everything else you have made there. From there you can download it or send it to a project. To make another, open it and click Open in Templates.

Virtual try-on and product holding

Open Templates in your account menu, then choose Virtual try-on or Product holding. These make an image without opening a project.

Choose a Person image, then add one to three references. Virtual try-on accepts garments and accessories; Product holding accepts product photos, packaging, other angles or companion items. Each image can come from an upload, a publicly accessible image URL, your Sandbox results or a project's generated images. To use a link, paste it into the image URL field and click Use image URL.

Instructions are optional. Describe which clothes to replace or keep, how to layer them, or which hand should hold a product and which way its logo should face. Under Advanced settings, choose an aspect ratio or leave Keep source ratio selected. The Seed starts at 5555; reuse it with the same images and settings to repeat a result.

Hover or focus Generate to see its estimated cost, then click it to make one image. Inputs stay locked while it runs. Cancel stops a queued or running job; cancellation is unavailable while the completed image is being saved. Your result replaces the example and is recorded in Your sandbox. Download image opens the output for saving.

Use these settings copies the example's instructions, aspect ratio and seed. It keeps your selected images, so it does not reproduce the example unless you supply the same sources. Template results cannot be replayed in the general Sandbox composer; open their template to make another.

Extend your video

Choose Extend your video to add footage at the End (the default) or Beginning. Upload an MP4, MOV, WebM or M4V, choose a saved Sandbox video, or select a generated shot or extension from one project. Imported local footage can be uploaded from your computer.

Sources must be no larger than 50 MB. End extensions accept 1.625–60 seconds of source and add 0.71–15 seconds. Beginning extensions accept 1–20 seconds of source and add 2–20 seconds. The form checks the source's readable video track before generation.

Write Continuation instructions describing the new action and camera movement. End extensions offer 480p, 768p, 1080p and 2K, with 768p selected initially. Under Advanced, choose Aspect ratio, turn Expand instructions on or off, or enter a Seed. The default aspect follows your source; changing it crops the video. Beginning extensions choose resolution automatically, preserve source aspect, and do not use these Advanced controls.

Hover, focus or press and hold Extend video to see the estimated price. End-extension estimates include a source-processing allowance. The final charge follows actual generated duration and source usage plus the usual fee; the added footage may be slightly longer than requested. Usage billing may remain pending briefly after the video arrives. Generation requires valid inputs and an available estimate. If usage billing is unavailable, end extension is refused before funds are reserved.

Play the completed result or choose Download video. When instructions were expanded, the result also shows Instructions used. Use these settings loads the example's settings while keeping your chosen source. Cancel stops queued or running work; completed outputs are retained while saving.

Was this page helpful?