# iTunes Library XML Parser Parse an iTunes or Music Library.xml in your browser and export playlists and track metadata as CSV, TSV, JSON, M3U or a library summary. ## Run it - **CLI:** `gizza tool itunes-library-parser ' Tracks ... Playlists ... '` - **Web:** https://gizza.ai/tools/itunes-library-parser/ - **Agents:** machine-readable descriptor (parameters JSON Schema) at https://gizza.ai/tools/itunes-library-parser/tool.json ## Inputs - `library` — Library.xml contents _(field)_ - `output` — Export as _(field)_ - `playlist` — Playlist name _(field)_ - `fields` — Columns _(field)_ - `path_prefix` — Replace music folder with _(field)_ - `path_style` — Path style _(field)_ - `include_builtin` — Include built-in playlists _(field)_ - `sort_by` — Sort rows by _(field)_ - `limit` — Row limit _(field)_ ## Output - Export (text) ## Query parameters Open the tool pre-filled and auto-run via URL: - `library` — Library.xml contents - `output` — Export as - `playlist` — Playlist name - `fields` — Columns - `path_prefix` — Replace music folder with - `path_style` — Path style - `include_builtin` — Include built-in playlists - `sort_by` — Sort rows by - `limit` — Row limit Example: `https://gizza.ai/tools/itunes-library-parser/?library=%3C%3Fxml%20version%3D%221.0%22%20encoding%3D%22UTF-8%22%3F%3E%0A%3Cplist%20version%3D%221.0%22%3E%0A%3Cdict%3E%0A%20%20%3Ckey%3ETracks%3C%2Fkey%3E%0A%20%20%3Cdict%3E...%3C%2Fdict%3E%0A%20%20%3Ckey%3EPlaylists%3C%2Fkey%3E%0A%20%20%3Carray%3E...%3C%2Farray%3E%0A%3C%2Fdict%3E%0A%3C%2Fplist%3E&output=csv&playlist=Late%20Night%20%28leave%20empty%20for%20every%20track%29&fields=name%2Cartist%2Calbum%2Cgenre%2Cyear%2Cduration%2Clocation&path_prefix=D%3A%5CMusic&path_style=original&include_builtin=true&sort_by=original` --- ## About this tool This parser reads the XML property list that Music and iTunes write as `Library.xml` and turns it into something you can actually use: a spreadsheet of your track metadata, a JSON dump for a script, an M3U playlist another player can open, an index of every playlist in the file, or a one-screen summary of the whole library. Everything runs on the text you paste. The library file is never uploaded, no music is read from disk, and nothing is written anywhere. That also means the tool works on a library exported from a machine you no longer have — all it needs is the XML. ## Getting your Library.xml In Music on macOS choose **File → Library → Export Library…** and save the XML file. In iTunes on Windows, the same command exists, and there is usually an `iTunes Music Library.xml` sitting next to your library in the iTunes folder (turn on the preference that shares the library XML with other applications if it is missing). Open the file in a text editor, copy all of it, and paste it above. The file must be the XML form — it starts with `` element. The binary `.itl` library next to it is a private database format and cannot be read by this tool or any other third-party one. ## Worked example A two-track library with one user playlist, exported with the default columns: ```xml Tracks 101 Track ID101 NameLet It Happen ArtistTame Impala AlbumCurrents GenreElectronic Year2015 Total Time467000 Locationfile://localhost/Users/me/Music/iTunes/iTunes%20Media/Music/Tame%20Impala/Currents/01%20Let%20It%20Happen.m4a ... ``` Output with `output=csv` and the default `fields`: ```text name,artist,album,genre,year,duration,location Let It Happen,Tame Impala,Currents,Electronic,2015,7:47,/Users/me/Music/iTunes/iTunes Media/Music/Tame Impala/Currents/01 Let It Happen.m4a All I Need,Air,Moon Safari,Electronic,1998,4:28,/Users/me/Music/iTunes/iTunes Media/Music/Air/Moon Safari/03 All I Need.m4a ``` Switching the same input to `output=m3u8` with `playlist=Late Night` gives an extended playlist in the playlist's own running order: ```text #EXTM3U #EXTINF:268,Air - All I Need /Users/me/Music/iTunes/iTunes Media/Music/Air/Moon Safari/03 All I Need.m4a #EXTINF:467,Tame Impala - Let It Happen /Users/me/Music/iTunes/iTunes Media/Music/Tame Impala/Currents/01 Let It Happen.m4a ``` ## Columns you can export `fields` takes a comma-separated list in the order you want the columns, and applies to the CSV, TSV and JSON outputs. The default is `name,artist,album,genre,year,duration,location`. Available: `track_id`, `name`, `artist`, `album_artist`, `composer`, `album`, `grouping`, `work`, `genre`, `kind`, `size`, `duration` (`m:ss`), `duration_seconds`, `disc_number`, `disc_count`, `track_number`, `track_count`, `year`, `bpm`, `date_added`, `date_modified`, `release_date`, `bit_rate`, `sample_rate`, `comments`, `play_count`, `play_date`, `skip_count`, `rating` (0–5 stars), `rating_raw` (the stored 0–100 value), `album_rating`, `loved`, `compilation`, `podcast`, `sort_name`, `sort_artist`, `sort_album`, `persistent_id`, `location`. JSON keeps native types, so a missing field is `null` and numbers stay numbers. CSV quotes any cell containing a comma, quote or newline; TSV strips tabs and newlines out of cells instead, because TSV has no quoting convention. ## Fixing file paths for another machine Track locations are stored as percent-encoded `file://` URLs, which are decoded back to ordinary paths for every output. Two options rewrite them: - **Replace music folder with** swaps the library's own music-folder prefix for a new root, so a playlist exported on one machine can point at a different drive letter, an external disk, or a network share. Tracks stored outside that folder keep their original path. - **Path style** controls slash direction: keep whatever was recorded, force forward slashes, force backslashes, or drop every folder and keep only the file name for a flat folder of copied music. ## Limits and edge cases - Up to 20 MB of library XML per run, which is roughly a 40,000-track library. - `limit` caps the number of track rows and is applied *after* sorting; `0` means no limit and the ceiling is 100,000 rows. - Only the XML property list form is readable; binary `.itl` libraries are not. - Tracks with no `Location` (cloud-only or missing files) come out with an empty path in CSV, TSV and JSON, and are skipped entirely in M3U, because a blank line is not a valid playlist entry. - Selecting a playlist *folder* returns an error rather than an empty export — folders hold other playlists, not tracks of their own. - Sorting is ascending and stable, using the library's sort-name fields when present; tracks missing the sort field are placed last rather than clustered at the top. - Smart playlists are exported from the track list the library recorded, not by re-evaluating their rules, so the result matches what the library last saved. ## FAQ
What is the difference between the two M3U outputs? `m3u` writes one file path per line and nothing else, which is what simple players and car stereos expect. `m3u8` writes the extended form: an `#EXTM3U` header plus an `#EXTINF` line before each path carrying the track's duration in seconds and an `Artist - Title` label. Both are UTF-8, so accented artist and album names survive either way.
Why don't I see the Music, Downloaded or Recently Added playlists? The playlist index hides the ones the app creates for itself so you only see your own. Turn on **Include built-in playlists** to list them too. They are detected from the `Master` and `Distinguished Kind` markers plus a list of well-known names, because libraries written by older versions and by third-party exporters often omit both markers. Exporting a built-in playlist by name always works regardless of that setting.
How do I find the exact playlist name to type? Run the tool once with **Export as → Playlist index**. It returns a CSV of every playlist with its name, kind (playlist, smart, folder or built-in), track count, parent folder and persistent ID. Playlist names are matched case-insensitively, so `late night` and `Late Night` both work, and an unrecognised name comes back with a list of the names that do exist in your file.
Why do the star ratings look odd? Ratings are stored as a 0–100 number in steps of 20, so three stars is `60`. The `rating` column converts that to 0–5 stars (half stars included), and `rating_raw` keeps the stored number if you would rather work with it directly. A track that has never been rated has no value at all and comes out blank, not as zero.
Can it copy or rename my actual music files? No. The tool only reads the pasted text and returns text, so it never touches your filesystem. What it can do is produce the paths a copy or move would need: use the file-name path style for a flat folder, or re-root the paths onto the drive you are copying to, then feed the resulting list to whatever tool actually moves the files.
## Related tools - [Podcast feed parser](https://gizza.ai/tools/podcast-feed-parser/): Parse a pasted podcast RSS/XML feed into clean JSON with episode titles, dates, durations, and audio enclosure URLs. Free, private, in-browser. - [Music File Renamer](https://gizza.ai/tools/music-file-renamer/): Build a safe rename and folder-move plan for music files from pasted tag dumps, templates, and filesystem rules. - [ARFF Converter](https://gizza.ai/tools/arff-converter/): Convert Weka ARFF datasets to CSV and CSV tables back to ARFF locally — nominal attributes, numeric types, dates, sparse rows, missing values, and type rows. - [Beancount to CSV Converter](https://gizza.ai/tools/beancount-to-csv/): Flatten Beancount or Ledger journal postings into spreadsheet-ready CSV, then rebuild a simple journal from that flat CSV schema. - [Convert BibTeX to CSV](https://gizza.ai/tools/bibtex-to-csv/): Paste BibTeX and get CSV, one row per entry. Choose columns, delimiter and header, decode LaTeX accents, reformat author names, sort rows and add an Excel BOM.