Built On Open Source
ProMedia runs entirely in your browser on the work of open-source communities. Every third-party component we send to your device is listed here, with its licence and where to find its source.
- 68 components
- 21 licence types
- Source mirror for GPL code
GNU GPL notice
FFmpeg engine — GPL-2.0-or-later
Some tools use an optional fallback engine, ffmpeg.wasm core 0.12.10: FFmpeg n5.1.4 compiled to WebAssembly with Emscripten 3.1.40, including x264, x265, libvpx, LAME, Opus, Vorbis, Theora, libwebp, FreeType, FriBidi, HarfBuzz, libass, zimg and zlib. It is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 2 of the License, or (at your option) any later version. It is distributed without any warranty; without even the implied warranty of merchantability or fitness for a particular purpose.
It is downloaded only when a tool needs it and runs as a separate program in its own Web Worker. ProMedia exchanges files and command lines with it through the MIT-licensed @ffmpeg/ffmpeg wrapper.
Served from:/engines/[email protected]/ffmpeg-core.js/engines/[email protected]/ffmpeg-core.wasm
What’s in the source mirror
FFmpeg & GPL codecs
- FFmpegn5.1.4GPL-2.0-or-later (as configured)
- x264 (ffmpegwasm fork)4-coresGPL-2.0-or-later
- x265 (ffmpegwasm fork)3.4GPL-2.0-or-later
Libraries linked into the core
- libvpx (ffmpegwasm fork)v1.13.1BSD-3-Clause
- LAME (ffmpegwasm fork)masterLGPL-2.0-or-later
- libogg (ffmpegwasm fork)v1.3.4BSD-3-Clause
- libtheora (ffmpegwasm fork)v1.1.1BSD-3-Clause
- libopus (ffmpegwasm fork)v1.3.1BSD-3-Clause
- libvorbis (ffmpegwasm fork)v1.3.3BSD-3-Clause
- libwebp (ffmpegwasm fork)v1.3.2BSD-3-Clause
- FreeType 2 (ffmpegwasm fork)VER-2-10-4FTL OR GPL-2.0-or-later
- GNU FriBidiv1.0.9LGPL-2.1-or-later
- HarfBuzz5.2.0MIT
- libass0.15.0ISC
- zimgrelease-3.0.5WTFPL
- zlib (ffmpegwasm fork)v1.2.11Zlib
Build scripts & toolchain
- ffmpeg.wasm (build scripts, bindings, Dockerfile)v12.15MIT
- Emscripten (toolchain; runtime glue + SDL2 port pins)3.1.40MIT OR NCSA
FFmpeg configure flags
--target-os=none --arch=x86_32 --enable-cross-compile --disable-asm --disable-stripping --disable-programs --disable-doc --disable-debug --disable-runtime-cpudetect --disable-autodetect --disable-pthreads --disable-w32threads --disable-os2threads --enable-gpl --enable-libx264 --enable-libx265 --enable-libvpx --enable-libmp3lame --enable-libtheora --enable-libvorbis --enable-libopus --enable-zlib --enable-libwebp --enable-libfreetype --enable-libfribidi --enable-libass --enable-libzimgLGPL components
These libraries are loaded as separate, unmodified files from our own server, only by the tools that need them. You may replace them with your own build; their source is in the source mirror.
heic-to (libheif + libde265, asm.js)
HEIC/HEIF decoding fallback, loaded as its own replaceable chunk.
LGPL-3.0@mediabunny/mp3-encoder (LAME 3.100)
MP3 encoding when the browser has no MP3 encoder.
MPL-2.0 + LGPL-2.0-or-later (LAME)@mediabunny/aac-encoder (FFmpeg AAC)
AAC encoding fallback (Firefox, Linux).
MPL-2.0 + LGPL-2.1-or-later (FFmpeg)
Third-party components
68 components
Framework & runtime
Astro7.3.8
Static site framework and client router.
React19.3.0
Interactive tool workspaces.
React DOM19.3.0
Tailwind CSS4.3.3
Utility CSS (compiled into the stylesheet).
Workbox7.4.1
Service worker: offline app shell and caching (precaching, routing, strategies, expiration).
Zustand5.0.15
Client state.
Interface
dnd kit (core)6.3.1
Accessible drag-and-drop reordering in merge tools.
dnd kit (sortable)10.0.0
dnd kit (utilities)3.2.2
Phosphor Icons2.1.1
Interface icons (SVG paths inlined at build).
three.js0.186.1
Hero light-beam scene (loaded lazily on large screens).
Fonts
Anton5.3.0
Heavy condensed type for meme captions and display watermarks (self-hosted via Fontsource). SIL Open Font License 1.1.
Inter Variable5.3.0
Interface typeface, self-hosted via Fontsource. SIL Open Font License 1.1.
JetBrains Mono5.3.0
Monospace option for text burned into videos (self-hosted via Fontsource). SIL Open Font License 1.1.
Nunito5.3.0
Rounded option for text burned into videos (self-hosted via Fontsource). SIL Open Font License 1.1.
Playfair Display5.3.0
Serif option for text burned into videos (self-hosted via Fontsource). SIL Open Font License 1.1.
Video & audio engines
@ffmpeg/util0.12.2
ffmpeg.wasm (@ffmpeg/ffmpeg)0.12.15
JavaScript wrapper and worker for the FFmpeg core.
Mediabunny1.61.3
Media demuxing/muxing and WebCodecs pipelines.
Engine libraries
FFmpeg (ffmpeg.wasm core)0.12.10
FFmpeg 5.1.4 compiled to WebAssembly with x264, x265, libvpx, LAME, libopus, libvorbis, libtheora, libwebp, libass, freetype, fribidi, harfbuzz, zimg, zlib. Served from /engines/[email protected]/; corresponding source offered per the deploy README. Downloaded only when a tool needs it.
FreeType
Inside the FFmpeg core.
FriBidi
Inside the FFmpeg core.
HarfBuzz
Inside the FFmpeg core.
Inter (subtitle font)
public/fonts/Inter-Regular.ttf, family name "Inter". Licence text: /fonts/Inter-OFL.txt.
libass
Subtitle rendering inside the FFmpeg core.
libavif + libaom
Inside @jsquash/avif.
libjxl
Inside @jsquash/jxl.
libvpx
Inside the FFmpeg core.
libwebp
Inside @jsquash/webp.
MozJPEG
Inside @jsquash/jpeg.
oxipng
Inside @jsquash/oxipng.
pako
Deflate support for UTIF.
x264
Inside the FFmpeg core.
Image
exifr7.1.3
EXIF/XMP/IPTC/GPS reading.
gifenc1.0.3
GIF encoding.
heic-to1.6.5
HEIC/HEIF decoding where the browser has none; separate lazily loaded, replaceable chunk.
imagetracerjs1.2.6
Raster-to-SVG tracing.
jSquash (MozJPEG, libwebp, libavif/aom, libjxl, OxiPNG, resize)1.6.0
WebAssembly image codecs, each loaded on first use.
LibRaw-Wasm1.6.0
RAW to JPG full decode, loaded only when needed. LibRaw is dual-licensed LGPL-2.1 / CDDL-1.0; its unmodified WebAssembly build runs as a separate lazily loaded module.
modern-gif2.1.0
GIF decoding where ImageDecoder is missing.
pdf-lib1.17.1
Images to PDF.
Tesseract.js7.0.0
On-device text recognition (OCR). Worker + LSTM cores (tesseract.js-core) are served from /engines/; language packs (tessdata_best int, Apache-2.0) download from jsDelivr after consent.
UTIF.js (utif2)4.1.0
TIFF decode/encode.
Audio
@mediabunny/aac-encoder1.61.3
Loaded only when the browser has no native AAC encoder.
@mediabunny/flac-encoder1.61.3
Loaded only for FLAC output.
@mediabunny/mp3-encoder1.61.3
Loaded only for MP3 output. MP3 encoding by the LAME project — https://lame.sourceforge.io
RNNoise (WASM)2025.1.5
WebAssembly build of RNNoise.
SoundTouchJS core2.1.1
WSOLA time-stretch for speed, pitch and voice effects (streamed). MPL-2.0: unmodified, source at the project URL.
AI runtime
@huggingface/tokenizers, @huggingface/jinja
Kokoro inference & text normalisation (ported from kokoro-js 1.2.1)
© hexgrad, Xenova. Vendored into src/engines/ai/worker/kokoro.ts to run on transformers.js v4.
ONNX Runtime Web1.31.0-dev.20260914-8d85527a0
WASM/WebGPU inference engine, self-hosted under /engines/ort@…/.
phonemizer 1.2.1 (eSpeak NG compiled to WebAssembly)
Text-to-phoneme step of text-to-speech. Bundles eSpeak NG (https://github.com/espeak-ng/espeak-ng), which is GPL-3.0-or-later; it is loaded only by the text-to-speech tools as a separate chunk. Source: the eSpeak NG repository; legal review recommended.
RNNoise (Xiph.Org / Jean-Marc Valin)
Transformers.js4.3.1
Model runtime (WebGPU / WASM), runs in a Web Worker.
AI models
all-MiniLM-L6-v2 (sentence-transformers) — Xenova ONNX export
Sentence embeddings for automatic chapters.
BiRefNet Lite (Zheng et al.) — onnx-community/BiRefNet_lite-ONNX
High-quality background removal.
Chrome built-in AI (Translator, LanguageDetector, Summarizer, Prompt API)
Used only when the browser provides it; runs on-device.
Kokoro-82M v1.0 (hexgrad) — onnx-community ONNX export
Text-to-speech model and voice style files.
MI-GAN (Picsart AI Research) — migan_pipeline_v2.onnx
Object and watermark removal (code and weights MIT).
MODNet (Ke et al.) — Xenova/modnet
Fast portrait background removal.
OPUS-MT translation models (Helsinki-NLP) — Xenova/opus-mt-* ONNX exports
Attribution: Jörg Tiedemann and Santhosh Thottingal, “OPUS-MT — Building open translation services for the World”, EAMT 2020.
Silero VAD v5
Self-hosted at /models/silero-vad/.
Swin2SR (Conde et al.) — classical ×2 and real-world ×4 (BSRGAN-PSNR)
Image super-resolution.
Whisper large-v3-turbo (OpenAI) — onnx-community export
Whisper tiny / base / small (OpenAI) — onnx-community *_timestamped exports
Speech recognition with word timestamps.
YuNet face detector 2023mar (OpenCV Zoo, Shiqi Yu et al.)
Self-hosted at /models/yunet/.
No components match your search.
Full licence texts
Copyright notices and licence texts of every package above, collected from the exact versions we ship.
Missing an attribution? Email [email protected] and we’ll fix it. ProMedia’s own code, design and brand are © 2026 Brainify.