Initializing... drag & drop files here
Supports: MKV
Matroska is the container people reach for precisely because it holds everything: several video tracks, a stack of audio languages, soft subtitles, chapters, attachments and tags. RFC 9559, published in October 2024, is its formal specification and covers Matroska versions 1 through 4.
A same-format conversion here does not preserve all of that. The job re-encodes the file and writes one video track and one audio track, plus at most one subtitle track. If your MKV is a multi-language rip with a commentary track and five subtitle streams, the output will be a shadow of it. That is the single most important thing to know about this page, and the reason the section below exists before the how-to rather than after it.
What a same-format pass is good for: changing the video codec inside the container, shrinking a file, dropping the resolution, forcing a batch onto one setting, or cutting a clip out of a long recording. Free, no sign-up, no watermark; uploads travel over an encrypted connection and are deleted automatically after a few hours.
Measured by pushing a purpose-built test file through the conversion with the page's own defaults — H.264 video, AAC audio, Matroska out.
| Element in the source | Count in | Count out | Note |
|---|---|---|---|
| Video track | 1 | 1 | Re-encoded, never copied |
| Audio tracks (eng, fra, deu) | 3 | 1 | Only the first survives; language tag carried over |
| Subtitle tracks (eng, fra) | 2 | 1 | Only the first survives |
| Subtitle format | SubRip | ASS | Text subtitles are rewritten into Matroska's native styled form |
| Chapters, attachments, tags | — | Do not rely on them | Extract anything you need before converting |
If you need every track intact, a converter is the wrong tool — nothing that re-encodes can promise that. Demux the extras with a Matroska tool locally, convert the video here, and re-mux afterwards.
| Codec | Why you would pick it | What to watch for |
|---|---|---|
| H.264 (default) | Universal hardware decode; safe for anything that will be played rather than archived | Larger than H.265 or AV1 at matched quality |
| H.265 | About 30-50% smaller than H.264 at the same visual quality | Patent-encumbered; not accepted by every service |
| AV1 | Royalty-free and the most efficient option offered | Slowest to encode by a wide margin |
| VP9 / VP8 | The WebM codecs, useful if the file will later be repackaged for the web | VP8 has no working CRF path; its Constant Quality control does nothing |
| FLAC, PCM (audio) | Lossless audio inside a lossy-video file — a common archival pattern | Grows the file substantially |
| MPEG-2 / MPEG-4 / Xvid / DivX | Compatibility with old hardware decoders | No CRF mode: Quality Preset becomes a multiple of the source bitrate, and the default asks for 132% of it |
| H.263 / H.261 | Almost never | These force a fixed frame size chosen from a short legal list based on your source dimensions. That overrides whatever you set under Video resolution and does not preserve the aspect ratio — a 1080p source is squashed into 704x576 |
| HuffYUV / MJPEG | Frame-accurate intermediate work | Enormous files; HuffYUV is lossless and hides most of the compression panel |
Because the conversion writes one audio stream. When a file with three language tracks goes through, only the first comes out — the others are not carried, not muted, simply absent. This is inherent to a re-encode: each output stream has to be encoded, and the pipeline builds a single video plus a single audio stream. Pull the extra tracks out with a local Matroska tool first if you need them.
No, it is Matroska's default text subtitle form. A SubRip track fed into a Matroska output is re-encoded into Advanced SubStation Alpha, which is a superset — the timings and text are intact and styling information is added. Every player that reads MKV subtitles reads ASS. If you specifically need the SRT file back, keep a copy before converting.
Yes. There is no stream-copy or remux path on our servers; the video is decoded to frames and encoded again on every job. At the default "Very High" preset on H.264 that resolves to CRF 18, which is close to visually lossless, but it is still a second generation of lossy compression. Do everything you need in one pass.
Because the Video Codec dropdown defaults to H.264, not to whatever your source used. H.264 needs roughly 30-50% more bitrate than H.265 for the same picture, so an HEVC-to-H.264 pass at a high quality preset commonly produces a larger file. Set Video Codec to H.265 explicitly if you want to stay in the same codec family, or accept the size increase in exchange for broader hardware compatibility.
Different default intent, and a different control set. This converter opens on Quality Preset and hides the "Target file size (%)" slider. Compress MKV does the reverse: it opens on that percentage slider and hides the preset. If your sentence is "make it 50% smaller", use the compressor. If it is "change the codec and drop it to 720p", use this page.
Not on this page — it writes video. MKV to MP3 does the extraction, re-encoding whichever audio track comes first into an MP3. Note the same one-track rule applies: if the film has an English track and a commentary track, you get the first one.
Because every job is a full re-encode, the cut lands where you ask rather than snapping to the nearest keyframe. That is the upside of re-encoding: frame-accurate in and out points, no black frames at the start, no audio drift. The downside is the encode time and the generation loss, which is the trade this whole page is built around.
No. The bitstream specification is open and is now published as an IETF standard, RFC 9559, in October 2024. It is built on EBML, a binary cousin of XML, which is what lets new element types be added without breaking older parsers. That openness is why MKV became the default container for archival rips and for anything that needs more track types than MP4 will hold.
Uploads and their outputs are deleted automatically after a few hours. Nothing is shared, published or indexed, and no account, email address or payment is required.