Wrong tier for clip length
“I pasted a 15s Reels clip but only got 5 seconds back.”
Aleph (Quick) processes only the first ~5s. Use Standard (Kling) for 3–15.5s social clips, or Pro (Seedance) for 4–30s full-length edits.
AI Video to Video
Edit Existing Clips — Change Look, Keep Motion
Paste TikTok / Reels / Facebook, or upload a clip — see common edit scenarios below, then try it in the tool. Same social import as Reference Studio.
Try the generator
Model
Source video to transform (Standard)
Paste TikTok / Reels / Facebook, or upload the clip — MP4 or MOV, 3–15.5 seconds. Motion and timing follow your upload; aspect ratio is automatic.
Paste a TikTok, Instagram Reels, or Facebook link — or a public HTTPS file URL ending in .mp4 or .mov. YouTube: upload MP4 only.
Style reference (optional)
Optional stills (up to 4) to steer look — works best with a clear transform prompt.
0/5000
Tip: describe what should change while keeping motion — e.g. "Transform into a rainy night scene while preserving the dancer's movements."
16:9 · 5s · 720p
Multimodal references, cinematic motion, and built-in audio in one generation.
Why video-to-video gets frustrating
Most overspending is picking the wrong tier or the wrong mode — not the model itself.
Background changed, subject drifted
“I only wanted a new background — the person looks different now.”
Write what to change and what to keep: "Replace background with neon city. Keep dancer motion and camera path."
Credits gone before a usable clip
“I ran Pro three times before I knew the look I wanted.”
Quick (Aleph) ~95 credits to test a look on ~5s, then rerun on Standard or Pro once motion is approved.
Popular video-to-video edit scenarios
Real jobs creators run on Reeldo — what you upload, what changes, and which tier fits.
Use case 01 / 06
Source
Edited
Product clip — swap background for A/B ads
- Scenario
- You have one product demo clip and need Amazon / Meta variants with different backgrounds — same motion, same framing.
- Result
- The product and camera path stay locked; only the environment changes for each campaign.
- Settings
- Standard (Kling)720p3–10s
Prompt template
Replace background with a clean white studio. Keep product shape, label readability, and camera static.
Use case 02 / 06
Source
Edited
Relight a clip — morning to golden hour
- Scenario
- Footage was shot in flat daylight but the ad brief calls for warm evening light — reshooting is not an option.
- Result
- Lighting and color grade shift while subject motion and timing stay on the original timeline.
- Settings
- Quick (Aleph) testthen Standard (Kling) final
Prompt template
Relight to golden hour with warm highlights. Preserve subject motion and framing.
Use case 03 / 06
Source
Edited
Remove a distraction in-frame
- Scenario
- A usable take has a passerby, logo, or prop that should not appear in the delivered ad.
- Result
- The unwanted element is gone; camera move and main subject action continue unchanged.
- Settings
- Standard (Kling)short prompt720p
Prompt template
Remove the person in the background on the left. Keep all other motion unchanged.
Use case 04 / 06
Source
Edited
Same UGC action — four looks for channel tests
- Scenario
- One talking-head or action clip must ship as TikTok, Reels, and paid social variants with different moods.
- Result
- Motion and timing match the source; only style, background, or grade changes per variant.
- Settings
- Standard (Kling)~16 cr/s @720p
Prompt template
Cyberpunk neon city background at night. Keep speaker gestures and lip timing.
Use case 05 / 06
Source
Edited
15–30s clip — edit without trimming
- Scenario
- Your source is a single 20s hero take and you cannot cut it down to Aleph or Kling limits.
- Result
- Output duration and aspect ratio follow the uploaded clip on Seedance Pro.
- Settings
- Pro (Seedance 2.5) only4–30s
Prompt template
Edit @Video1: cinematic teal-orange grade. Keep all motion and duration.
Use case 06 / 06
Source
Edited
Image-guided look — packaging or color ref
- Scenario
- Marketing sent a still of the approved packaging or palette; the video must match without a reshoot.
- Result
- The clip adopts colors or product details from the reference still while motion follows the source video.
- Settings
- Pro (Seedance) @image1or Kling + up to 4 refs
Prompt template
Apply colors and packaging style from @image1 to @Video1. Keep camera motion.
Pick Quick, Standard, or Pro
Route by clip length and budget — not by brand hype.
| Your clip | File / import | Budget mindset | Pick |
|---|---|---|---|
| ≤5s style test | ≤10MB upload | Lowest cost (~95 cr/job) | Quick — Aleph |
| 3–15s Reels / ads | TikTok / Reels / FB or upload | ~16 cr/s @720p | Standard — Kling |
| 4–30s cannot trim | Upload or social import | ~48 cr/s @720p Pro | Pro — Seedance 2.5 |
How to write a video-to-video edit prompt
Tell the model what to change and what must stay locked on the original timeline.
- →Keep motion — "preserve dancer moves / camera path / lip timing"
- →Change look — background, lighting, color grade, or season
- →Seedance Pro — "Edit @Video1: …" and optional @image1 reference still
- →Kling Standard — optional up to 4 reference stills for look steering
Example
Replace background with rainy neon street. Keep dancer motion and camera path.
Short, direct prompts work best on Aleph. Add @Video1 on Seedance Pro for long clips.
Supported inputs & Kie limits
Same social import as Reference Studio — tier duration limits still apply after import.
| Input | Supported? | Notes |
|---|---|---|
| Upload MP4 / MOV / WebM | Yes | Aleph ≤10MB · Kling & Seedance ≤200MB |
| TikTok / Instagram Reels / Facebook | Yes | Sign in · wait for import to finish · same as Reference Studio |
| Direct HTTPS .mp4 / .mov URL | Yes | Public CDN or signed object URL Kie can fetch |
| YouTube / Shorts links | No (site-wide) | Download a short MP4 and upload instead |
| After social import | Tier limits apply | 20s Reels → Kling or Seedance · not Aleph (5s) |
How AI video to video works
Step 1
1. Add your source clip
Paste TikTok / Reels / Facebook, or upload MP4/MOV. Match tier limits — Aleph: ≤10MB · Kling: 3–15.5s · Seedance: 4–30s.
Step 2
2. Pick Quick, Standard, or Pro
Use the model dropdown in the generator. Standard (Kling) is the default for most social edits.
Step 3
3. Describe the edit
Say what should change — background, lighting, objects, or style — while keeping the original motion.
Step 4
4. Generate & download
Preview the result, iterate on another tier if needed, then download. Credits scale with length on Standard and Pro.
When not to use Video to Video
V2V edits footage you already have in place — other modes handle generation and extensions.
No source clip yet — start from a prompt or still
Text to Video →Generate a new scene from refs — not edit this timeline
Reference to Video →Add seconds at the end without changing look
Video Extend →Sharpen resolution only — no style edit
Video Upscale →Compare AI video to video tiers
| Feature | Quick (Aleph) | Standard (Kling) | Pro (Seedance 2.5) |
|---|---|---|---|
| Best for | Style tests & quick fixes | Social posts & ad cutdowns | Long clips & final delivery |
| Clip length | ~5s processed (10MB max) | 3–15.5s (200MB) | 4–30s, matches source |
| Resolution | 720p equivalent | 720p · 1080p · 4K | 480p draft · 720p · 1080p · 4K |
| Typical cost | ~95 credits / job | ~16 cr/s @720p | ~48 cr/s @720p (input+output) |
| Prompt tip | Short & direct: "Replace background with neon city." | Preserve motion: "Rainy night scene, keep dancer moves." | Use tags: "Edit @Video1: golden hour cinematic look." |
Video to video vs other modes
Video to video edits existing footage in place — unlike image or reference modes that start from stills or mixed refs.
| Feature | Video to Video | Image to Video | Reference to Video |
|---|---|---|---|
| Input | One source video + edit prompt | Start frame image + motion prompt | Reference images, videos, and audio |
| Goal | Change an existing clip in place | Animate a still into a new clip | Generate a new scene guided by references |
| Motion | Preserves source camera & subject motion | Motion invented from the still | Motion can follow reference clips |
| Duration | Aleph ~5s · Kling 3–15.5s · Seedance 4–30s | Model picker (typically 4–30s) | Up to 30s on Seedance 2.5 |
After you generate
Your edited clip previews in the generator above. From there you can download, or continue in Reeldo:
- Download MP4 when the clip is ready.
- Video Extend — add seconds at the end in the same scene.
- Video Upscale — sharpen resolution before social upload or delivery.