Batch upscaling is where the credit math really matters: one photo is cheap, fifty photos is a budget decision. This guide covers when batching makes sense, how to prepare files so a bad one does not waste a batch, the two batch paths on SeedVR2.net, and the per-image costs you should know before uploading.
When batching makes sense
Batch upscaling fits three situations well:
- Product catalogs that need consistent resolution across every SKU
- Event shoots or family scans where every file needs the same treatment
- Archive projects where files are similar in size and quality
If your files vary wildly — mixed resolutions, some sharp and some blurry — sort them first. Batching assumes one target fits all.
Prep: sort and check before uploading
Spend ten minutes so the batch runs clean:
- Delete near-duplicates and obviously broken files
- Run the important ones through the free resolution check to see each file's true dimensions
- Group files by target: everything within roughly 1.5x of 2K goes to 2K; only large originals go to 4K
- Confirm the formats are supported — JPEG, PNG, and WebP are accepted
- Estimate the total with the upscale cost calculator before you commit
The cheapest mistake to avoid: upscaling a tiny, blurry file to 4K and paying 16 credits for invented detail. If a source is too small for its target, reshoot or rescan it instead.
Homepage batch workflow
The SeedVR2 workspace on the homepage accepts multiple images in one job — in its current configuration, up to 1,000 images per batch with a 2 GB total size cap. The uploader shows its live limits, so trust what the interface says over any summary.
The workflow:
- Drag a folder of prepared images into the homepage workspace
- Pick one target resolution for the batch — 2K (10 credits each), 4K (16), or 8K (25)
- Confirm the total in the preflight
- Submit and download the finished set
Because every image in the job shares one resolution, mixed batches mean some files get more upscale than they need. Group by target first.
Photo-enhance batches
When photos need improvement, not just size — soft focus, grain, old compression damage — use the photo enhance tool. It accepts up to 10 files per batch and offers two modes:
- Faithful: 150 credits per image, JPEG output. Sharpens while staying close to the original
- Reconstruction: 90 credits per image, PNG output. Attempts deeper repair on degraded files
Photo enhance is a different job than resolution upscaling, and it costs accordingly: a full 10-photo batch is 1,500 credits in Faithful mode or 900 in Reconstruction. That makes the mode choice the real budget decision — test one photo in each mode before committing a batch of ten.
Credit math for realistic batches
| Batch | Cost |
|---|---|
| 10 product photos to 2K | 100 credits |
| 10 product photos to 4K | 160 credits |
| 10 photos, photo enhance (Reconstruction) | 900 credits |
| 10 photos, photo enhance (Faithful) | 1,500 credits |
A new account's free first 2K image is a practical quality check before you spend on a batch. Guests can preview results at 720px before unlocking, and current pack prices are on the pricing page.
Common batch mistakes
Four mistakes account for most wasted batch credits:
- Mixed resolutions in one job. A batch shares one target, so a 600px thumbnail and a 3000px hero cannot both upscale well at 2K. Sort first.
- Double sharpening. If you sharpen in an editor and then run photo enhance, halos appear around edges. Sharpen once, at the end.
- Wrong tool for the job. Pure size problems belong in the image upscaler (10 to 25 credits). Damage and softness belong in photo enhance (90 to 150 credits). Running enhance on a clean, large photo mostly burns credits.
- Skipping the spot check. One bad render in a batch of fifty goes unnoticed until a customer zooms in.
After the batch
Do not trust a batch blindly:
- Spot-check every output at 100% zoom
- Compare against originals for color shifts or oversharpening halos
- Rename outputs clearly so originals and upscales cannot be mixed up
- Re-check any marketplace or print requirements against the finished files
Batching multiplies both your time savings and your mistakes — the prep checklist is what keeps it profitable. Start with the free first 2K on the image upscaler, then queue real batches once you know the pipeline handles your files well.