Remove AI Metadata from Multiple Images at Once: A Batch Workflow

How to strip AI metadata from a batch of JPEG, PNG, or WebP files in one pass. Sequential processing, per-file download, no upload, no account required.

Sep 21, 2026

AI Metadata Ninja Team

If you publish more than one image at a time — a product line, a content calendar, a print catalog — the single-file workflow becomes tedious. You want to drop the whole folder in, see what each file carries, and download a clean version of each. This guide covers the batch workflow: what the drop zone does with multiple files, how the daily quota is consumed, and where the batch breaks down.

For the single-file case, the one-image guide walks through the same flow with one file. The verb-variant hub covers remove / clean / delete / strip / scrub / erase in one place.

How batch processing actually works in the drop zone

There is no separate "batch mode" in the tool. You drop multiple files into the same drop zone, and the scanner processes them in sequence. The drop zone accepts up to the daily quota per drop (10 anonymous, 20 signed-in, or your credit balance for paid users).

For each file, the scanner produces a report. The reports are stacked vertically under the drop zone, in the order the files finish scanning. Each report has its own Clean button and its own Download button. There is no "clean all" button — you decide per file whether to clean, and you can clean them in any order.

Sequential processing is the bottleneck: 10 files take roughly 10x the time of one file. The byte-level walk is fast (under a second per file on a modern laptop), so 10 files finishes in under 10 seconds, and 20 files in under 20 seconds. If you have hundreds of files, see the paid credit packs — the per-credit cost is about $0.05 for the smallest pack, dropping to $0.033 for the bulk pack.

What changes when you drop a batch

Three things, all visible in the UI:

Quota counter ticks down per file. Drop 10 files at once, and the counter goes from 0/10 to 10/10 as each file is scanned. The 11th file, if you drop it, is rejected with "today's free limit is 10 files."
Reports stack under the drop zone. Each report shows the file name, the AI fields found, and the Clean button. Files with no AI metadata show "no AI fields found" — you do not need to clean them, but doing so produces a byte-identical output.
Downloads are per file. There is no "download all as zip" button. Each file is downloaded individually as the Clean completes. If you close the page mid-batch, the files you have already cleaned are still on your machine.

Sequential processing is intentional. Concurrent processing would multiply the memory footprint (each file loaded as ArrayBuffer) and could exceed browser limits on lower-end devices. The bottleneck is not the CPU — it is the I/O.

When the batch makes sense

The batch flow is the right choice when:

You are publishing a set of images — a product line, a content calendar, a portfolio update. The cleaning policy is the same for all of them: strip everything AI-related.
You are migrating a corpus — moving an old library of AI-generated images into a clean state. The 10-a-day free tier is too slow for this, but a 500-credit pack at $20 covers a meaningful chunk.
You are running a recurring clean — every Monday, clean the new images from the week. The free tier covers the small case; the credit packs cover the larger one.

The single-file flow is the right choice when:

Each image needs a different decision — some keep the C2PA credential for audit, some strip it. You cannot make that decision in batch without a UI for per-file policy.
You are working on a sensitive file — one image that needs extra care. The single-file flow gives you more space to read the report and decide.
You are below quota and the file is the last one for the day. Drop it alone so the quota counter is not accidentally consumed by a misclicked batch.

Workflow for a typical batch

Gather the files in a single folder. JPEGs, PNGs, and WebPs can be mixed. HEIC files are read-only and should be converted first.
Drop the folder into the drop zone. The browser expands the folder and adds each file to the queue. (Drag a folder from Finder or Explorer; the drop zone accepts folders as well as individual files.)
Wait for the reports. Each report appears under the drop zone as the scan finishes. The total time scales linearly with the number of files.
Click Clean on each report. The cleaner runs in the background. You can clean them in any order; each Clean button is independent.
Download the cleaned files. Each cleaned file downloads as it is ready. Save them into a "cleaned" subfolder, or replace the originals in place.
Verify the batch. Spot-check two or three cleaned files with exiftool (or drag them back into the drop zone). The XMP parameters, EXIF Software, and C2PA blocks should be empty.

Daily limits, in concrete terms

| Tier | Per day | Per month equivalent | Best for | | ------------------- | -------- | -------------------- | -------------------------------- | | Anonymous | 10 scans | ~300 scans | Trying the tool, one-off batches | | Signed-in (free) | 20 scans | ~600 scans | Light recurring use | | Starter pack ($5) | — | 100 scans | One batch cleanup, no recurring | | Standard pack ($20) | — | 500 scans | Mid-volume recurring use | | Bulk pack ($50) | — | 1500 scans | High-volume or agency work |

The paid credits do not expire. Buy a pack when you have a batch to run, and use them as needed.

When the batch flow breaks down

Two cases where you should drop back to single-file:

HEIC files. The scanner reads them and reports what is inside, but the cleaner cannot write HEIC. Convert to JPEG first (macOS Preview can do this in a batch action), then drop the JPEGs in.
Files over 10 MB. The drop zone rejects files over 10 MB. The limit is a browser memory cap, not a tool choice. If you have larger images, downscale them first or split the metadata segments differently.

Everything else — JPEG, PNG, WebP, mixed formats, mixed sizes — works in batch.