FlashVSR Video Upscaler
FlashVSR is a video super-resolution model that upscales an existing clip and sharpens its detail without changing its content, length or frame rate
What is FlashVSR?
FlashVSR is a video super-resolution model that upscales an existing clip and sharpens its detail without changing its content, length or frame rate. On blubiu it takes one MP4 or MOV up to 120 seconds and outputs 720P, 1080P, 2K or 4K, billed per second of the source video. It runs in the browser on blubiu's servers, so there is nothing to install.
Checked 2026-11-02
What FlashVSR is good at
- +Old or archive footage that looks soft on modern screens.
- +Social video that lost detail to platform or messaging-app compression.
- +Keeping an AI clip exactly as it is, such as a 480P take from Seedance 2.0 Mini or Wan 3.0, and lifting it to 1080P–4K for delivery.
- +Finishing character videos made on blubiu: a 720P motion transfer or a 720P recast goes to 2K or 4K without running the generation again.
- +Preparing an editing master: lift a 1080P clip to 2K before you crop or reframe it in your editor.
When not to use it
- −The source is so blurry that the detail never existed. FlashVSR sharpens what is there; if a new take is acceptable, regenerate the shot at 1080P with Seedance 2.0. See Seedance 2.0 →
- −You want to change the content, the style or the person in the clip. FlashVSR only changes resolution; Wan 3.0 can restyle from references or replace the person. See Wan 3.0 →
- −You want 2K straight from a prompt rather than from existing footage. MiniMax H3 renders 2K directly. See MiniMax H3 →
- −You want more frames or slow motion. FlashVSR keeps the source frame rate and doesn't interpolate.
FlashVSR versions and specs
| Version | Inputs | Length | Resolution | Sound | Price |
|---|---|---|---|---|---|
| FlashVSR | One video (up to 120 s) | Same as your video (up to 120 s) | 720P / 1080P / 2K / 4K | — | 720P 4 · 1080P 6 · 2K 7 · 4K 12 credits per second of your video; subscribers 40% off: 720P 2.4 · 1080P 3.6 · 2K 4.2 · 4K 7.2 credits/s |
Standard prices, read live from the platform catalog; once signed in, the Generate button shows your account's actual price. Failed generations are refunded.
How to prompt FlashVSR
Start from the cleanest source you have
There's no prompt: the source file is the only input. Upload the original rather than a screen recording or a copy saved back from a chat app, because every round of compression removes detail the upscaler can't get back.
Trim before you upload
You pay per second of the source, so cut it to the part you'll use. Clips over 120 seconds need to be split first.
Match the target to the screen
Pick the resolution the video will actually play at. A higher tier costs more per second and can't add detail beyond what the source holds.
Re-export files with no duration
Some streamed or fragmented MP4 recordings don't store a total length, and the page can't price them. Re-export as a regular MP4 and upload again.
Upload one clip, then pick the tier
The form at the top of this page is the same one as the AI video upscaler: upload one clip, pick 1080P, 2K or 4K, and the Generate button shows the total before you confirm.
Upscale only the takes you keep
Choose takes and trim them at the original resolution first, then send only what made the cut. Check faces and on-screen text at full size before you swap the upscaled file into your edit.
FlashVSR examples
All generated on blubiu, unedited.
FlashVSR FAQ
What is FlashVSR?
FlashVSR is a video super-resolution (VSR) model: it upscales an existing video and sharpens detail without generating new content. On blubiu you run it in the browser with credits, right on this page or from the AI video upscaler, with no local setup.
How is FlashVSR priced?
Per second of the source video, at a rate set by the target resolution. The price is fixed when you submit, so the size of the output file doesn't change it. Current rates are in the table above, the Generate button shows the total, and failed jobs are refunded automatically.
Can I try FlashVSR for free?
Yes, with the Free plan's monthly credits. A short clip at 1080P is the cheapest way to see what it does with your footage, the Generate button shows the exact cost first, and failed jobs are refunded. Plans and credit packs are on the pricing page.
How long can the video be, and what changes?
Up to 120 seconds per clip, as MP4 or MOV. The output keeps the source duration and frame rate; only resolution and sharpness change.
Which resolution should I choose?
The one the video will play at: 1080P for everyday viewing, 2K for an editing master, 4K for large screens or delivery. A higher tier costs more per second and can't add detail the source doesn't hold, so start from the cleanest copy you have.
Can FlashVSR fix a very blurry video?
Only partly. It makes edges and textures sharper, but it can't recover faces, text or detail the camera never captured. The cleaner the source, the more the upscale shows.
Can I upscale AI-generated clips?
Yes, and it is a common use. Draft a shot cheaply with text to video or image to video, keep the take you like, and FlashVSR lifts it to 1080P–4K with the same motion and timing. A new generation at a higher tier would be a different clip.
FlashVSR or regenerating with a video model?
Use FlashVSR to keep the exact clip you have, with the same motion, timing and framing. Regenerate with Seedance 2.0 or Wan 3.0 at 1080P when a new take is acceptable; it will be a different clip each time.
Do I need to install FlashVSR or run it locally?
No. On blubiu it runs on our servers: upload the clip in the browser and save the result when it's done. You don't need a GPU, ComfyUI or any local setup.
Is this page different from the AI video upscaler?
Same model and the same upload form, so you can run it from either page. The AI video upscaler page is about the job (when upscaling helps and which tier to pick); this page is about the model: what FlashVSR is, where it falls short and what it costs.
Tools that use FlashVSR
Pair FlashVSR with
AI Image to Video Generator →
Animate a still at 480p or 720p to test the shot cheaply, then upscale only the take you keep.
AI Text to Video Generator →
Prompt-only clips can be drafted at a low tier and finished at 1080P–4K without regenerating them.
AI Motion Transfer →
Motion transfer renders 720P or 1080P; take a finished dance clip to 2K or 4K here.
Recast Video with AI →
Recasts default to 720P; upscale the keeper instead of running the swap again at a higher tier.
Other video models
Make 4–15 second videos with sound from text, a first frame or up to 9 reference images with ByteDance's Seedance 2.0 — Mini, Fast and 2.0 compared.
Use MiniMax H3, the Hailuo video model, for 2K clips with reference images and videos, or H3 Lite for the cheapest silent drafts from 1 second.
Generate an 8-second 720p clip with optional sound using Google's Veo 3.1 Fast, from text or from a first and last frame. One fixed price per clip.
Use Alibaba's Wan 3.0 for 2–15 second videos with sound from text, a first frame or reference images, or to replace the person in a video.
Run Kling Motion Control 2.6 online: a full-body photo copies the moves from a 3–30 second video, at 720P or 1080P. No Kling account needed; try it free.
See image models →