How an MP3 stores tags and artwork
An MP3 file stores metadata defined by the ID3v2.4 specification.
Modern tags sit at the start of the file and hold named frames: TIT2 for title, TPE1 for
artist, TALB for album, and APIC for a picture. Older files may carry ID3v1
instead — a fixed 128-byte block at the very end, with hard character limits and a numeric genre code.
An APIC frame contains a complete image file — a whole JPEG or PNG, byte for byte — plus a
MIME type and a picture-type code saying whether it is a front cover, a back cover or a band logo.
Extracting artwork is therefore a copy, not a conversion, which is why there is no quality loss.
How to extract album art from an MP3
- Open the track. Drop the .mp3 onto the box above, or click to browse. Only the first few megabytes are read, so the file opens instantly regardless of its length.
- Look at the artwork. Every embedded image is shown with its type, format, pixel dimensions and file size. Most tracks carry one front cover; some carry a back cover and a disc image as well.
- Save the image. Click Save image. The bytes written are the bytes embedded in the file, so nothing is re-compressed.
- Read and copy the tags. Every readable ID3 field is listed underneath. Copy tags puts the whole set on your clipboard as label-and-value lines.
What comes out
- Embedded artwork, saved byte-for-byte at original resolution, with its picture type identified — front cover, back cover, leaflet, band logo and so on.
- Every readable text tag — title, artist, album, album artist, track, disc, year, genre, composer, publisher, BPM, ISRC, copyright, comments and encoder.
- The tag version in use, which tells you what software wrote the file.
- ID3v1 fallback. If there is no ID3v2 block, the 128-byte v1 tag at the end of the file is read instead, with its numeric genre code translated to a name.
Where both tag versions exist, ID3v2 wins and v1 only fills in fields v2 does not have — which is the same precedence media players use.
What is supported, and what is not
ID3v2.2, v2.3 and v2.4 are all read, along with ID3v1 and v1.1. Two details commonly break simpler parsers and are handled here: ID3v2 sizes are stored "syncsafe", using only seven bits per byte so tag data can never be mistaken for audio, and ID3v2.2 uses three-character frame ids with three-byte sizes instead of four.
Text is decoded according to each frame's own encoding byte — Latin-1, UTF-16 with either byte order, or UTF-8 — so non-English titles read correctly rather than as mojibake. Only the first 8 MB is read, which covers any realistic tag block including large embedded artwork.
M4A, AAC, FLAC, OGG and WAV files store metadata in completely different containers and are not read here. They are detected and reported rather than misparsed.
Why the track is never uploaded
The tag block is read from your disk by the browser and parsed there. No server takes part, so the audio is never transmitted — and because only the first few megabytes are read at all, most of the file is never even loaded into memory.
The practical benefit is speed as much as privacy: uploading a 90-minute set to read its cover art would take minutes, and reading the header takes milliseconds.
What this tool does not do
This is a reader, not a tag editor. Four boundaries:
- It does not write or change tags. Use Mp3tag, MusicBrainz Picard or foobar2000 for that.
- It does not find missing artwork. If a track has no
APICframe there is nothing to extract — the artwork you see in your player is being fetched from an online database, not read from the file. - It does not read audio properties. Bitrate, sample rate and duration come from the audio frames, not the tag block.
- It does not read other audio formats. FLAC uses Vorbis comments, M4A uses MP4 atoms; neither is ID3.
Lyrics stored in USLT frames are read, but synchronised lyrics in SYLT are
not.
Why people extract MP3 artwork and tags
- Recovering cover art — pulling a high-resolution sleeve out of a track when the original image is long gone.
- Building a music library — checking what a file actually claims before importing it.
- Diagnosing a player — finding out why a track shows the wrong artist, which is almost always a stale ID3v1 tag disagreeing with ID3v2.
- DJs and radio — reading ISRC, BPM and publisher fields for reporting.
- Archiving — recording the metadata of a collection before re-tagging it.
Album art compared with other embedded images
Artwork in an MP3 behaves exactly like an icon inside a Windows executable or a picture inside a Word document: a complete image file stored whole inside a larger container, extractable without loss. The container differs; the principle does not.
For pictures inside a Word, Excel or PowerPoint file, use the Office image extractor. For icons inside a program, use the icon extractor. To read the text pictured in an image rather than the image itself, use image to text OCR.
For complete walkthroughs and technical specifications, read our master guide on how to extract album art and ID3 metadata from MP3 files, or learn the binary architecture in how album art is stored in MP3 files.
ID3 tag format standards and edge cases
ID3v2.3 and ID3v2.4 are the two versions in common use; the extractor reads both.
An ID3v2 tag is a 10-byte header (ID3, version, flags, and a size in four
synchsafe integers), followed by frames. Each frame has a four-character ID: TIT2 (title),
TPE1 (artist), TALB (album), TRCK (track number) and APIC (attached picture) are the most
common.
Three edge cases: (a) the APIC frame begins with an encoding byte, a MIME type string
(e.g., image/jpeg), a picture type byte (3 = front cover, 4 = back cover,
0 = other) and a description string before the image binary — a file with two APIC frames
both typed as front cover (type 3) is unusual but valid, and both are returned; (b) synchsafe
integers use only the lower 7 bits of each byte, making a 4-byte size field hold at most
268,435,455 bytes rather than 4,294,967,295; (c) ID3v1 tags — the 128-byte fixed-length
block at the end of the file — use Latin-1 encoding with no charset declaration, so
non-ASCII artist or title characters may display incorrectly when no ID3v2 tag is also
present.
Frequently asked questions
How do I extract album art from an MP3?
Drop the file onto this page. Every embedded image appears with a Save image button, at full original resolution.
Will the cover image lose quality?
No. The image is stored whole inside the tag, so extraction copies the exact bytes. There is no re-encoding.
Is my music file uploaded to a server?
No. Only the tag block at the start of the file is read, and it is parsed inside your browser.
Why does my track show no album art?
It has no APIC frame. Players often display artwork fetched from an online database rather than stored in the file, which makes it look embedded when it is not.
Can I edit tags or add an album cover with this?
No — this tool reads, it does not write. To put a cover onto an MP3, use Mp3tag (Windows), MusicBrainz Picard (any system) or foobar2000. Extract the image here first if you are copying artwork from another track.
What is the difference between album art and an album cover?
Nothing — they are the same thing, stored in the ID3 APIC frame. Players and tools use the terms interchangeably, along with cover art and embedded artwork.
Why do the artist and album look wrong?
The file probably has both an ID3v1 and an ID3v2 tag that disagree. This tool shows ID3v2 first, which is what most players use, and fills gaps from v1.
Does it work on FLAC, M4A or OGG?
No. Those formats store metadata in different containers — Vorbis comments and MP4 atoms — not ID3. The tool says so rather than guessing.
Why are non-English titles readable here but not elsewhere?
Each ID3 frame declares its own text encoding. This tool honours that declaration, so UTF-16 and UTF-8 titles decode correctly instead of appearing as garbled characters.