How to change id3 tags without uploading — free browser tool
- Step 1Load the editor before going offline (optional) — Open the ID3 editor. Because processing is local, you can disconnect from the network after the page loads and still tag files.
- Step 2Drop the sensitive file — Drop one recording. It is read into the tab's memory only. There is no server-side copy and no multi-file batch upload.
- Step 3Verify no upload in the network panel — Open DevTools → Network, then process. Confirm no request carries your audio. This is the privacy guarantee made observable.
- Step 4Enter the metadata you want recorded — Fill Title, Artist (interviewer/source), Album (case/project name), Year, Genre, and Track # as needed. Blank fields keep any existing tag.
- Step 5Embed a cover only if appropriate — Optionally add a JPG/PNG (e.g. a case sheet). Be aware embedded art is visible to anyone who opens the file — omit it for truly sensitive material.
- Step 6Download and store the tagged copy — Process and save the new
*-taggedfile directly to your secure location. The original is untouched; nothing was transmitted.
The six editable fields and the ID3 frames they write
These are the exact fields the editor exposes. Each non-empty value is passed to FFmpeg as -metadata key=value; the value is normalised by FFmpeg's MP3 muxer into the matching ID3v2.3 frame. Empty fields are skipped entirely, so the original frame is preserved by the audio stream-copy.
| Editor field | FFmpeg metadata key | ID3v2.3 frame | Notes |
|---|---|---|---|
| Title | title | TIT2 | Track / song / episode title |
| Artist | artist | TPE1 | Lead performer / author / host |
| Album | album | TALB | Album / show name / collection |
| Year | date | TYER / TDRC | FFmpeg writes the Year box to the date key, not a separate year key |
| Genre | genre | TCON | Free text — no fixed picklist is enforced |
| Track # | track | TRCK | Plain number or n/total (e.g. 3/12) — typed as text |
Tier limits for the ID3 editor
ID3 editing stream-copies the audio (no re-encode), so it is one of the lightest audio operations — but the same per-file ceilings apply. Tagging is single-file: the tool does not accept a multi-file drop, so the per-batch file count does not apply here.
| Tier | Max file size | Max duration | Files per run | Cost |
|---|---|---|---|---|
| Free | 50 MB | 30 minutes | 1 (single file) | Free, no account |
| Pro | 200 MB | 120 minutes | 1 (single file) | Paid |
| Pro-media | 100 GB | Unlimited | 1 (single file) | Paid |
| Developer | 100 GB | Unlimited | 1 (single file) | Paid |
Input format → output format and metadata model
The editor keeps the output in the same container as the input — there is no format selector. The same six fields map onto each container's native metadata model. The MP3 muxer is the only one that gets the explicit -id3v2_version 3 flag.
| Input | Output | Metadata model | Album art |
|---|---|---|---|
.mp3 | .mp3 | ID3v2.3 (forced via -id3v2_version 3) | APIC attached-picture frame |
.m4a / .mp4 (AAC/ALAC) | same | iTunes-style MP4 metadata atoms | covr atom |
.flac | .flac | Vorbis comments | FLAC picture block |
Cookbook
Tagging patterns for confidential audio, with the privacy behaviour made explicit. Identifying details are placeholders.
Label an interview without revealing the source online
A journalist tags an interview recording with an internal case code rather than a name, and does it locally so the file is never uploaded to a third-party tagger.
Fields: Title = Interview 2026-06-03 (Source B) Artist = Newsroom / J.D. Album = Case 'Riverside' Year = 2026 Network panel during processing: 0 requests carrying audio Output: interview-tagged.mp3 saved locally only
Tag on an air-gapped machine
On a restricted laptop where uploads are blocked by policy, the page is loaded once on the corporate network, then the file is tagged offline because FFmpeg runs in WASM.
1. Load /audio-tools/id3-editor while online 2. Disconnect network 3. Drop confidential.mp3, fill fields, process 4. Download tagged file -- all offline (works because no server round-trip is required)
Re-tag without re-encoding evidence audio
For a chain-of-custody clip, re-encoding would alter the bytes. The stream-copy guarantees the audio samples are unchanged — only the container's metadata header is rewritten.
ffprobe before: codec=mp3 bit_rate=192000 duration=612.3 ffprobe after : codec=mp3 bit_rate=192000 duration=612.3 Audio stream identical; only ID3 header differs
Strip a misleading old tag by overwriting it
A file arrived with someone else's name in the Artist tag. Type the correct value to overwrite it; the field you fill replaces the old frame.
Before: Artist=Wrong Name Title=clip You type: Artist = Confidential After: TPE1=Confidential (Title preserved via -c copy) Note: to fully clear a field, overwrite it -- a blank box keeps the old value
Project-code an HR recording set
An investigator tags each clip with a consistent Album (the matter number) and Track # so the set stays ordered, processing each one locally.
clip-1.mp3 -> Album=HR-2026-014 Track=1/4 clip-2.mp3 -> Album=HR-2026-014 Track=2/4 (single file per run; no batch upload) All processing in-browser, nothing transmitted
Edge cases and what actually happens
No audio leaves the browser
Private by designFFmpeg runs in WebAssembly in the tab. There is no upload of the audio and no server-side temporary file. You can confirm via the network panel.
A blank field does not erase the old tag
PreservedLeaving a box empty keeps the existing frame. To remove a misleading value you must overwrite it with the correct text — there is no per-field 'clear' that blanks a tag.
Embedded cover art is visible to anyone
ExpectedAn APIC cover is not hidden — anyone who opens the file sees it. For sensitive material, do not embed an identifying image.
Filename itself can leak information
Out of scopeThe tool tags the file's metadata, not its name. A revealing filename (e.g. a source's name) is still on disk — rename the file separately.
Page reload before download loses the result
ExpectedBecause everything is in memory, reloading the tab discards the processed file. Download before navigating away.
File exceeds the tier duration limit
413 over limitA long deposition recording may pass the size cap but fail the duration cap (Free 30 min, Pro 120 min). Split it with the audio-splitter or upgrade.
Audio is never re-encoded
Lossless by design-c copy means the recording's samples are unchanged — important when the audio is evidence and must not be altered.
Works offline after load
SupportedOnce the page has loaded its WASM assets, tagging works with the network disconnected. There is no licence check or telemetry needed to process.
Non-image file dropped as cover
Rejected (invalid)The cover picker accepts only JPG/PNG. A PDF or document cannot be embedded as artwork.
Frequently asked questions
Does my recording get uploaded anywhere?
No. FFmpeg 8.1 runs in your browser via WebAssembly. The audio is read into the tab, tagged locally, and handed back. No server receives the file — verify it in the network panel.
Can I use this offline?
Yes, after the page loads. Processing is local, so you can disconnect from the network and still tag files. This is ideal for restricted or air-gapped machines.
Is the audio altered when I change tags?
No. The audio stream is copied with -c copy — never re-encoded. Only the metadata header changes, which matters for chain-of-custody or evidence audio.
Do I need an account to keep it private?
No account is required on the Free tier. Not signing in means nothing ties the file to an identity on our side — though processing is local regardless.
How do I remove a wrong name from a tag?
Overwrite the field with the correct value. A blank box preserves the old tag, so you must type the replacement text to change it.
Can someone recover the original from the tagged file?
The tagged file contains the same audio (stream-copied) and your new metadata. It is a normal audio file — protect it like any other sensitive document. The tool adds no hidden recovery data.
Is the filename private too?
The tool changes metadata, not the filename. If the filename itself is sensitive, rename the file separately before sharing.
Can I tag an interview that's two hours long?
On Pro you can (120-minute cap); the Free tier caps at 30 minutes. Long files can also be split first with the audio-splitter.
Will embedding cover art expose anything?
Embedded art is plainly visible to anyone who opens the file. For confidential audio, leave the cover empty.
Does it work on M4A and FLAC, or only MP3?
All three. M4A/MP4 uses iTunes metadata atoms, FLAC uses Vorbis comments, MP3 uses ID3v2.3 — the same six fields apply, all processed locally.
Is there any telemetry or analytics on the file?
The processing pipeline does not transmit your audio. The tool needs no network call to tag a file once loaded.
What happens to the file when I close the tab?
It is discarded from memory. Nothing persists server-side because nothing was ever sent.
Privacy first
Every JAD Audio tool runs entirely in your browser via FFmpeg (WebAssembly) and RNNoise. Your audio files never leave your device — verified by zero outbound network requests during processing.