MKVEXTRACT(1)

📦
version 102.0 · 15 versions
, in fedora 44

💖 MKVEXTRACT(1) — user commands, but make it a vibe 💖

okay so bestie, welcome to the mkvextract(1) manual page 🫶 this is the User Commands corner of the manual and it's literally about pulling stuff OUT of Matroska™ files. sit down, hydrate, let's go ✨

✨ NAME ✨

mkvextract — extract tracks from Matroska(TM) files into other files 💅

🌸 SYNOPSIS 🌸

mkvextract {source-filename} {mode1} [options] [extraction-spec1]
           [mode2] [options] [extraction-spec2] [...]

🦋 DESCRIPTION 🦋

So this program extracts specific parts out of a Matroska(TM) file and turns them into other useful formats. The very first argument is the name of the source file, and that file must be a Matroska(TM) file, no cap. 🎀

Every other argument either:

  • switches to a certain extraction mode,
  • changes options for the mode you're currently in, or
  • says what to extract into which file.

And here's the iconic part: multiple modes can be used in the same invocation of mkvextract, so you can extract a whole bunch of things in a single pass 👑 Most options only work in certain modes though — a few are universal besties that apply everywhere.

Currently supported is the extraction of tracks, tags, attachments, chapters, CUE sheets, timestamps and cues. slay.

💫 Common options 💫

These ones are available in all modes, so we only explain them once, right here, mwah 💋

  • -f, --parse-fully — sets the parse mode to 'full'. The default mode does not parse the whole file; instead it uses the meta seek elements to locate the required elements of a source file. In 99% of all cases that's honestly enough. But for files that don't contain meta seek elements, or that are damaged, you might have to use this mode. Heads up: a full scan of a file can take a couple of minutes, while a fast scan only takes seconds 😭
  • --command-line-charset character-set — sets the character set to convert strings given on the command line from. It defaults to the character set given by system's current locale. 🌍
  • --output-charset character-set — sets the character set to which strings are converted that are to be output. Also defaults to the character set given by system's current locale. 💬
  • -r, --redirect-output file-name — writes all messages to the file file-name instead of to the console. You could just do output redirection, sure, but there are cases where you genuinely need this option: when the terminal reinterprets the output before writing it to a file. The character set set with --output-charset is honored. ✨
  • --no-bom — normally text files that are created & encoded in one of the UTF variants will start with a byte order marker (BOM). Using this option disables writing that byte order marker. bye BOM 👋
  • --flush-on-close — tells the program to flush all data cached in memory to storage when closing files opened for writing. This can be used to prevent data loss on power outages, or to circumvent certain problems in the operating system or drivers. The downside? Multiplexing will take longer, because mkvmerge will wait until all data has been written to the storage before exiting. See issues #2469 and #2480 on the MKVToolNix bug tracker for in-depth discussions on the pros and cons. 🔥
  • --ui-language code — forces the translations for the language code to be used (e.g. 'de_DE' for the German translations). Entering 'list' as the code will cause the program to output a list of available translations. 🗣️
  • --abort-on-warnings — tells the program to abort after the first warning is emitted. The program's exit code will be 1. 💀
  • --debug topic — turn on debugging for a specific feature. This option is only useful for developers. ngl if you're not a dev, skip it 🙈
  • --engage feature — turn on experimental features. A list of available features can be requested with mkvextract --engage list. These features are not meant to be used in normal situations. 🚨
  • --gui-mode — turns on GUI mode. In this mode specially-formatted lines may be output that can tell a controlling GUI what's happening. These messages follow the format '#GUI#message'. The message may be followed by key/value pairs as in '#GUI#message#key1=value1#key2=value2...'. Neither the messages nor the keys are ever translated and always output in English. 👑
  • -v, --verbose — be verbose and show all the important Matroska(TM) elements as they're read. 🍓
  • -h, --help — show usage information and exit.
  • -V, --version — show version information and exit.
  • @options-file.json — reads additional command line arguments from the file options-file. For a full explanation on the supported formats for such files see the section called "Option files" in the mkvmerge(1) man page. 📄

🎬 Track extraction mode 🎬

Syntax:

mkvextract source-filename tracks [options] TID1:dest-filename1
           [TID2:dest-filename2 ...]

The following command line options are available for each track in the 'tracks' extraction mode. They have to appear in front of the track specification (see below) they should be applied to. so like, order matters bestie, don't fight it 💅

  • -c character-set — sets the character set to convert the next text subtitle track to. Only valid if the next track ID targets a text subtitle track. It defaults to UTF-8. ✨
  • --blockadd level — keep only the BlockAdditions up to this level. The default is to keep all levels. This option only affects certain kinds of codecs like WAVPACK4. 🎧
  • --cuesheet — causes mkvextract(1) to extract a CUE sheet from the chapter information and tag data for the following track into a file whose name is the track's output name with '.cue' appended to it. 🎀
  • --raw — extracts the raw data into a file without any container data around it. Unlike the --fullraw flag this flag does not cause the contents of the CodecPrivate element to be written to the file. This mode works with all CodecIDs, even the ones that mkvextract(1) doesn't support otherwise, but the resulting files might not be usable. ⚠️
  • --fullraw — extracts the raw data into a file without any container data around it. The contents of the CodecPrivate element will be written to the file first if the track contains such a header element. This mode works with all CodecIDs, even the ones that mkvextract(1) doesn't support otherwise, but the resulting files might not be usable. ⚠️
  • TID:outname — causes extraction of the track with the ID TID into the file outname if such a track exists in the source file. This option can be given multiple times. The track IDs are the same as the ones output by mkvmerge(1)'s --identify option. 🆔

Each output name should be used only once. The exception are RealAudio and RealVideo tracks. If you use the same name for different tracks then those tracks will be saved in the same file. Example:

$ mkvextract input.mkv tracks 0:video.h264 2:output-two-vobsub-tracks.idx 3:output-two-vobsub-tracks.idx

📎 Attachments extraction mode 📎

Syntax:

mkvextract source-filename attachments [options] AID1:outname1
           [AID2:outname2 ...]
  • AID:outname — causes extraction of the attachment with the ID AID into the file outname if such an attachment exists in the source file. If the outname is left empty then the name of the attachment inside the source Matroska(TM) file is used instead. This option can be given multiple times. The attachment IDs are the same as the ones output by mkvmerge(1)'s --identify option. 💌

📖 Chapters extraction mode 📖

Syntax:

mkvextract source-filename chapters [options] output-filename.xml
  • -s, --simple — exports the chapter information in the simple format used in the OGM tools (CHAPTER01=..., CHAPTER01NAME=...). In this mode some information has to be discarded. Default is to output the chapters in XML format. 📝
  • --simple-language language — if the simple format is enabled then mkvextract(1) will only output a single entry for each chapter atom encountered, even if a chapter atom contains more than one chapter name. By default mkvextract(1) will use the first chapter name found for each atom, regardless of its language. 🌐

Using this option allows the user to determine which chapter names are output if atoms contain more than one chapter name. The language parameter must be an ISO 639-1 or ISO 639-2 code.

The chapters are written to specified output file. By default the XML format understood by mkvmerge(1) is used. If no chapters are found in the file, the output file is not created. 🚫

🏷️ Tags extraction mode 🏷️

Syntax:

mkvextract source-filename tags [options] output-filename.xml
  • -T, --no-track-tags — exports only tags that aren't track-specific. 🙅♀️

The tags are written to specified output file in the XML format understood by mkvmerge(1). If no tags are found in the file, the output file is not created. 🚫

💿 Cue sheet extraction mode 💿

Syntax:

mkvextract source-filename cuesheet [options] output-filename.cue

The cue sheet is written to specified output file. If no chapters or tags are found in the file, the output file is not created. 🚫

⏱️ Timestamp extraction mode ⏱️

Syntax:

mkvextract source-filename timestamps_v2 [options] TID1:dest-filename1
           [TID2:dest-filename2 ...]
  • TID:outname — causes extraction of the timestamps for the track with the ID TID into the file outname if such a track exists in the source file. This option can be given multiple times. The track IDs are the same as the ones output by mkvmerge(1)'s --identify option. 🆔

Example:

$ mkvextract input.mkv timestamps_v2 1:ts-track1.txt 2:ts-track2.txt

🎯 Cues extraction mode 🎯

Syntax:

mkvextract source-filename cues [options] TID1:dest-filename1
           [TID2:dest-filename2 ...]
  • TID:dest-filename — causes extraction of the cues for the track with the ID TID into the file outname if such a track exists in the source file. This option can be given multiple times. The track IDs are the same as the ones output by mkvmerge(1)'s --identify option and not the numbers contained in the CueTrack element. 👀

The format output is a simple text format: one line per CuePoint element with key=value pairs. If an optional element is not present in a CuePoint (e.g. CueDuration) then a dash will be output as the value. ➖

Example:

timestamp=00:00:13.305000000 duration=- cluster_position=757741 relative_position=11

The possible keys are:

  • timestamp — the cue point's timestamp with nanosecond precision. The format is HH:MM:SS.nnnnnnnnn. This element is always set. ⭐
  • duration — the cue point's duration with nanosecond precision. The format is HH:MM:SS.nnnnnnnnn. ⏳
  • cluster_position — the absolute position in bytes inside the Matroska(TM) file where the cluster containing the referenced element starts. 📍
Note ✨ Inside the Matroska(TM) file the CueClusterPosition is relative to the segment's data start offset. The value output by mkvextract(1)'s cue extraction mode, however, contains that offset already and is an absolute offset from the beginning of the file. iykyk 💅
  • relative_position — the relative position in bytes inside the cluster where the BlockGroup or SimpleBlock element that the cue point refers to starts.
Note ✨ Inside the Matroska(TM) file the CueRelativePosition is relative to the cluster's data start offset. The value output by mkvextract(1)'s cue extraction mode, however, is relative to the cluster's ID. The absolute position inside the file can be calculated by adding cluster_position and relative_position. math queen behavior 🧮💖

Example:

$ mkvextract input.mkv cues 1:cues-track1.txt 2:cues-track2.txt

🌈 EXAMPLES 🌈

Extracting both chapters and tags in their respective XML formats at the same time, bestie:

$ mkvextract movie.mkv chapters movie-chapters.xml tags movie-tags.xml

Extracting a couple of tracks and their respective timestamps at the same time — multitasking icon behavior 👑:

$ mkvextract "Another Movie.mkv" tracks 0:video.h265 "1:main audio.aac" "2:director's comments.aac" timestamps_v2 "0:timestamps video.txt" "1:timestamps main audio.txt" "2:timestamps director's comments.txt"

Extracting chapters in the Ogg/OGM format and re-encoding a text subtitle track to another character set:

$ mkvextract "My Movie.mkv" chapters --simple "My Chapters.txt" tracks -c MS-ANSI "2:My Subtitles.srt"

🎀 TEXT FILES AND CHARACTER SET CONVERSIONS 🎀

For an in-depth discussion about how all tools in the MKVToolNix suite handle character set conversions, input/output encoding, command line encoding and console encoding, please see the identically-named section in the mkvmerge(1) man page. go read it, it's giving thorough 💜

🦋 OUTPUT FILE FORMATS 🦋

The decision about the output format is based on the track type, not on the extension used for the output file name. Like, the extension is NOT the vibe check here, the track type is 🫶 The following track types are supported at the moment:

  • A_AAC/MPEG2/*, A_AAC/MPEG4/*, A_AAC — all AAC files will be written into an AAC file with ADTS headers before each packet. The ADTS headers will not contain the deprecated emphasis field.
  • A_AC3, A_EAC3 — these will be extracted to raw AC-3 files.
  • A_ALAC — ALAC tracks are written to CAF files.
  • A_DTS — these will be extracted to raw DTS files.
  • A_FLAC — FLAC tracks are written to raw FLAC files.
  • A_MPEG/L2 — MPEG-1 Audio Layer II streams will be extracted to raw MP2 files.
  • A_MPEG/L3 — these will be extracted to raw MP3 files.
  • A_OPUS — Opus(TM) tracks are written to OggOpus(TM) files.
  • A_PCM/INT/LIT, A_PCM/INT/BIG — raw PCM data will be written to a WAV file. Big-endian integer data will be converted to little-endian data in the process. glow up conversion fr fr ✨
  • A_REAL/* — RealAudio(TM) tracks are written to RealMedia(TM) files.
  • A_TRUEHD, A_MLP — these will be extracted to raw TrueHD/MLP files.
  • A_TTA1 — TrueAudio(TM) tracks are written to TTA files. Please note that due to Matroska(TM)'s limited timestamp precision the extracted file's header will be different regarding two fields: data_length (the total number of samples in the file) and the CRC. heads up, not a bug, just the precision situation 🙈
  • A_VORBIS — Vorbis audio will be written into an OggVorbis(TM) file.
  • A_WAVPACK4 — WavPack(TM) tracks are written to WV files.
  • S_HDMV/PGS — PGS subtitles will be written as SUP files.
  • S_HDMV/TEXTST — TextST subtitles will be written as a special file format invented for mkvmerge(1) and mkvextract(1). bespoke, we made it ourselves 💅
  • S_KATE — Kate(TM) streams will be written within an Ogg(TM) container.
  • S_TEXT/SSA, S_TEXT/ASS, S_SSA, S_ASS — SSA and ASS text subtitles will be written as SSA/ASS files respectively.
  • S_TEXT/UTF8, S_TEXT/ASCII — simple text subtitles will be written as SRT files.
  • S_VOBSUB — VobSub(TM) subtitles will be written as SUB files along with the respective index files, as IDX files.
  • S_TEXT/USF — USF text subtitles will be written as USF files.
  • S_TEXT/WEBVTT — WebVTT text subtitles will be written as WebVTT files.
  • V_MPEG1, V_MPEG2 — MPEG-1 and MPEG-2 video tracks will be written as MPEG elementary streams.
  • V_MPEG4/ISO/AVC — H.264 / AVC video tracks are written to H.264 elementary streams which can be processed further with e.g. MP4Box(TM) from the GPAC(TM) package.
  • V_MPEG4/ISO/HEVC — H.265 / HEVC video tracks are written to H.265 elementary streams which can be processed further with e.g. MP4Box(TM) from the GPAC(TM) package.
  • V_MS/VFW/FOURCC — fixed FPS video tracks with this CodecID are written to AVI files.
  • V_REAL/* — RealVideo(TM) tracks are written to RealMedia(TM) files.
  • V_THEORA — Theora(TM) streams will be written within an Ogg(TM) container
  • V_VP8, V_VP9 — VP8 / VP9 tracks are written to IVF files.
  • Tags — tags are converted to a XML format. This format is the same that mkvmerge(1) supports for reading tags.
  • Attachments — attachments are written to the output file as they are. No conversion whatsoever is done. untouched, no notes, periodt 💅
  • Chapters — chapters are converted to a XML format. This format is the same that mkvmerge(1) supports for reading chapters. Alternatively a stripped-down version can be output in the simple OGM style format.
  • Timestamps — timestamps are first sorted and then output as a timestamp v2 format compliant file ready to be fed to mkvmerge(1). The extraction to other formats (v1, v3 and v4) is not supported. sorry bestie, v2 only 😭

💀 EXIT CODES 💀

mkvextract(1) exits with one of three exit codes:

  1. 0 — this exit code means that extraction has completed successfully. slay 🎉
  2. 1 — in this case mkvextract(1) has output at least one warning, but extraction did continue. A warning is prefixed with the text 'Warning:'. Depending on the issues involved the resulting files might be ok or not. The user is urged to check both the warning and the resulting files. read the warning, check ur files, don't skip this step 🙈
  3. 2 — this exit code is used after an error occurred. mkvextract(1) aborts right after outputting the error message. Error messages range from wrong command line arguments over read/write errors to broken files.

🌸 ENVIRONMENT VARIABLES 🌸

mkvextract(1) uses the default variables that determine the system's locale (e.g. LANG and the LC_* family). Additional variables:

  • MKVEXTRACT_DEBUG, MKVTOOLNIX_DEBUG and its short form MTX_DEBUG — the content is treated as if it had been passed via the --debug option.
  • MKVEXTRACT_ENGAGE, MKVTOOLNIX_ENGAGE and its short form MTX_ENGAGE — the content is treated as if it had been passed via the --engage option.

💜 SEE ALSO 💜

mkvmerge(1), mkvinfo(1), mkvpropedit(1), mkvtoolnix-gui(1)

🌈 WWW 🌈

The latest version can always be found at the MKVToolNix homepage[1].

👑 AUTHOR 👑

Moritz Bunkus <moritz@bunkus.org> Developer

🎀 NOTES 🎀

  1. the MKVToolNix homepage https://mkvtoolnix.download/

and that's the whole page bestie 💖 this one's from MKVToolNix 102.0, dated 2026-09-14, and it's the MKVEXTRACT(1) manual. go extract something iconic ✨🦋

📦 where this page comes from