MLT Multimedia Framework

Author, manage, and run multitrack audio/video compositions.
The engine of a non-linear video editor that can be used in all sorts of apps, not just desktop video editors.
MLT is an open source multimedia framework, designed and developed for television broadcasting. It provides a toolkit for broadcasters, video editors, media players, transcoders, web streamers and many more types of applications. The functionality of the system is provided via an assortment of ready to use tools, XML authoring components, and an extensible plug-in based API.
Download source code. The easiest way to try out and learn MLT is by downloading Shotcut.

Recent News

WebVfx v1.1.0 released

Jun 15, 2019

You can download it from GitHub.

MLT continues to maintain and extend the WebVfx add-on that lets one use HTML and Qt QML technologies for a video asset or effects. Back in November, 2018 we released v1.0.0 as it has proven fairly stable and useful in Shotcut for a few years. Here are the enhancements for v1.1.0:

  • Added support for alpha channels in MLT video for input and output.
  • Added a transparent property to the MLT filter. When the property is set to 1, this gives a filter a transparent white canvas for drawing while the MLT video image is still available through the JavaScript extension. This facilitates using WebVfx to draw the alpha channel independently of the MLT alpha channel.
  • Added support for getting a MLT rectangle property (mlt_rect) as a JavaScript object with properties: x, y, width, height, and opacity.

v6.16.0 released

May 7, 2019

You can download it from GitHub.

This version is released to facilitate packaging the latest version of Shotcut, which is using new APIs.


Added functions to get/set a creation date to a producer:

  • mlt_producer_get_creation_time()
  • mlt_producer_set_creation_time()
  • Mlt::Producer::set_creation_time()
  • Mlt::Producer::get_creation_time()


  • Fixed dance filter not showing when lower track is transparent.
  • Refactored dynamictext filter to use mlt_producer_get_creation_time().
  • Marked frei0r rgsplit0r plugin version < 1.1 as not thread-safe.
  • Fixed possible null pointer crash in mlt_properties_serialise_yaml().

v6.14.0 released

Mar 30, 2019

You can download it from GitHub.

This version is mostly fixes plus a few API additions and filters.


  • Added mlt_profile_lumas_dir().
  • Added mlt_frame_get_unique_properties().
  • Added mlt_playlist_reorder() and Mlt::Playlist::reorder().
  • Added some new convenience constructors to mlt++
    • Producer(mlt_profile profile, const char *id, const char *service = NULL)
    • Consumer(mlt_profile profile, const char *id , const char *service = NULL)
    • Transition(mlt_profile profile, const char *id, const char *arg = NULL)
    • Filter(mlt_profile profile, const char *id, const char *service = NULL)
    • Tractor(mlt_profile profile, char *id, char *arg = NULL)
  • Added Mlt::Transition::connect(Service&).
  • Added unit tests for mlt_playlist.
  • Fixed a crash on invalid transition track values in mlt_transition.
  • Fixed a deadlock regression in v6.12.0 of mlt_consumer when starting from a paused state (producer speed=0).


  • The avformat module now requires at least FFmpeg v2.4 or Libav 12.
  • Added mask_start and mask_apply filters to the core module.
  • Added qtext filter to qt module.
  • Changed dynamictext and timer filters to use qtext.
  • Fixed number of digits for seconds in timer filter.
  • Added mlt_image_format property to color producer.
  • Improved color accuracy of libswscale RGB->YUV conversion.
  • Fixed frei0r producers not working with tractor.
  • Fixed decklink consumer stalling on dropped frames.
  • Generate lumas for 16:9, 9:16 (vertical), and square aspect ratios.
  • Fixed crash in qimage when alpha_size is zero.
  • Fixed the mlt_consumer channels property not being passed to multi consumer.
  • Fixed the shape filter for full range color and crashes.
  • Converted the shape filter to use mlt_animation.
  • Added a use_mix property to the shape filter.
  • Fixed invert=1 and mix=100 gives wrong image in shape filter.
  • Fixed a possible free null pointer in the linsys sdi consumer.
  • Fixed using destroyed temporary object in qimage.
  • Fixed a possible null pointer dereference in the spot_remover filter.
  • Fixed memory leak on swr_convert() failure in swresample filter.
  • Fixed possible null pointer dereference in affine when not using rect.
  • Fixed loading image sequence on Windows in qimage.
  • Fixed some null pointer crashes using Movit opengl services.
  • Fixed sdl2 consumer crashes during initialization on Linux or BSD.
  • Fixed distorted image using melt_file.
  • Fixed qimage build on Qt version < 5.5.
  • Added offset property to the timer filter.
  • Changed the boxblur hori & vert properties’ minimum to 0.
  • Fixed crash in duplicate frame on rotated videos.
  • Added automatic scaling and padding to avfilter.
  • Fixed field order when encoding progressive as interlace.
  • Fixed frei0r plugins to use the number of slices from the threads property.
  • Fixed over compositing with transparent clips in luma transition.
  • Added sliced processing to dissolve-with-alpha using the threads property.
  • Added createdate keyword to dynamictext filter.
  • Fixed possible crash changing audio_index in avformat producer.
  • Fixed small memory leaks in xml consumer, jackrack, and timewarp producer.
  • Fixed compiling opencv module with OpenCV > 3.


  • Added vertical video profiles:
    • vertical_hd_30
    • vertical_hd_60
  • Mlt++ now requires C++11 compiler support.
  • Added --disable-windeploy to configure to keep bin & lib folders on Windows.
  • Added support for consumer in & out to melt.
  • Fixed color accuracy of lossless/Ut Video preset and use pix_fmt yuv422p.
  • Fixed lossless/H.264 preset to use crf=0.
  • Fixed compiling with mingw32.
  • Fixed build with Python 3.

v6.12.0 released

Nov 26, 2018

You can download it from GitHub.

This version has many important fixes plus a few new filters and support for encoding using VA-API.


  • Changed buffer property to be mutable and adaptive to speed property in mlt_consumer.
  • Changed macOS RELOCATABLE build to use standard app bundle layout:
    • lib/mlt -> ../PlugIns/mlt
    • lib/frei0r-1 -> ../PlugIns/frei0r-1
    • lib/ladspa -> ../PlugIns/ladspa
    • share/mlt -> ../Resources/mlt
    • share/movit -> ../Resources/movit
  • Fixed a_track of transitions matching deleted track in mlt_tractor_remove_track().
  • Fixed multi-thread race crash in mlt_properties_clear().
  • Fixed possiblle null pointer crash in mlt_property_get_rect() and mlt_property_get_time().
  • Fixed non-animated strings containing ‘;’ or ‘=’ in mlt_animation_parse().
  • Fixed crash in clear_property() with mlt_animation.


  • Added a generic text filter to the plus module.
  • Added a timer filter to the plus module.
  • Added audio timeout handling to sdl2 consumers.
  • Added spot_remove filter to the plus module.
  • Added dds, ico, and webp filename extensions for qimage producer.
  • Added support for color_range property in avformat consumer: “pc” or “jpeg” for full range, otherwise limited range.
  • Added a window property to the audiowaveform filter.
  • Added MM:SS.SS to the timer filter.
  • Added query string param “multi” to the xml producer to force using the multi consumer.
  • Improved WebP image support in avformat producer.
  • Integrated hwupload filter in avformat consumer if using VAAPI codec.
  • Changed count producer to use pango if qtext not available.
  • Changed qt moduled to not call XInitThreads()
  • Changed color producer to only set alpha on frame if rgb24a requested or not opaque.
  • Changed the xml producer to pass quality and performance parameters to the multi consumer.
  • Fixed sdl2_audio distortion (regression in v6.10.0).
  • Fixed dynamictext filter to not error on empty text.
  • Fixed dynamictext aliased (regression in v6.10.0).
  • Fixed qimage outputs premultiplied if scaled internally.
  • Fixed crash in cbrts consumer if running property was never set.
  • Fixed rendering edges of some typefaces in qtext producer.
  • Fixed qimage fails to load with wrong filename extension.
  • Fixed affine dark right and bottom edge artifacts regression in (v6.10.0).
  • Fixed support for vp8 and vp9 with alpha channel in avformat producer.
  • Fixed interpolation mode selection in qimage producer.
  • Fixed crash in qimage with alpha channel.
  • Fixed some AAC MP4 files start playing from middle in avformat producer.
  • Fixed crash in avfilter if initialization fails.
  • Fixed crash in mix when frame rate is very low.
  • Fixed crash on missing luma file in composite transition.
  • Fixed A/V sync on some files in avformat producer.
  • Fixed seeking on audio filter with album art in avformat producer.
  • Fixed colorspace conversion in avformat consumer.


  • Added more avformat consumer presets:
    • alpha/Quicktime Animation
    • alpha/vp8
    • alpha/vp9
    • alpha/Ut Video
    • lossless/Ut Video
  • Added square video profiles:
    • square_1080p_30
    • square_1080p_60
  • Added support for nodejs to the swig bindings.
  • Changed configure script to require opencv module be explicitly enabled.
  • Numerous spelling fixes in source code and comments thanks to codespell.

v6.10.0 released

Jul 2, 2018

You can download it from GitHub.

This version fixes bugs and supports serializing animation keyframes with a specified time format (previously only frame number).


  • Reverted mlt_pool change in v6.8.0 pending further testing.
    (USE_MLT_POOL compiler define is now a 0/1 boolean, defaults to 1.)
  • Fixed crash regression in v6.8.0 “parsing non-animated string as an animation.”
  • Added pointer checks to mlt_animation.
  • Changed producer cache size heuristic in mlt_multitrack to be more liberal.
  • Fixed handling reserved characters in names for YAML in mlt_properties.
  • Added clamping to prevent computing negative in and out points to mlt_producer.
  • Added functions to serialize animation with a time format:
    • mlt_animation_serialize_cut_tf()
    • mlt_animation_serialize_tf()
    • mlt_property_get_string_tf()
    • mlt_property_get_string_l_tf()
    • mlt_properties_get_value_tf()
    • Mlt::Properties::get(int, mlt_time_format)
    • Mlt::Animation::serialize_cut(mlt_time_format, int, int)
  • Added functions to clear a property to mlt_properties:
    • mlt_property_clear()
    • mlt_properties_clear()
    • Mlt::Properties::clear()


  • Fixed enabling sliced pix_fmt conversion in avformat producer.
  • Fixed incorrect seek and sync on audio files with discard packets.
  • Added support for avcodec_send_frame() API to avformat consumer.
  • Fixed compile errors with Libav master.
  • Fixed a crash in affine transition.
  • Fixed a crash in ladspa filters when consumer frame rate is low (e.g. <= 8).
  • Fixed a crash in boxblur filter.
  • Added animation support to boxblur hori and vert properties.
  • Fixed a crash in movit.convert.
  • Fixed incorrect alpha in affine transition blending routine.
  • Converted frei0r from deprecated mlt_geometry to mlt_animation API.
  • Fixed tilde in text string for pango producer.
  • Fixed using more than one channelcopy filter.
  • Fixed the mono filter reducing volume level.
  • Fixed degraded audio scrubbing in sdl2_audio consumer.
  • Converted dynamictext filter to use affine transition for more correct alpha compositing and sub-pixel positioning.
  • Added time format support for animation keyframes to the xml consumer.
  • Added animation support to more affine transition properties:
    • fix_rotate_x
    • fix_rotate_y
    • fix_rotate_z
    • fix_shear_x
    • fix_shear_y
    • fix_shear_z
    • ox
    • oy
    • scale_x
    • scale_y
  • Fixed gaps in text when characters overlap in qtext and kdenlive producers.
  • Fixed a crash in pixbuf producer with multiple render threads.
  • Converted the oldfilm vignette filter from mlt_geometry to mlt_animation.


  • Numerous updates to mlt-xml.dtd.
  • Categorized many of the encode presets (using


Featured Apps






Subscribe to Newsvia RSS.


MLT enables you to author, manage, and run multitrack audio/video compositions.
See our Hall of Fame
Copyright © 2008-2018 by Meltytech, LLC.

Social Links