No Upload Studio
Licences
No Upload Studio is built on open-source code and free fonts made by other people. Here is every part, its licence and where its source lives.
Code
-
Mediabunny 1.61.0
Reads and writes video and audio files in the video and audio editors.
- Licence
- Mozilla Public License 2.0
- Source
- github.com/Vanilagy/mediabunny
- Copyright
- Vanilagy and contributors
Each editor that uses it carries a build of the unchanged source with only the parts it needs, with its notice at the top of that code.
-
lamejs 1.2.1
Saves MP3 files in the audio editor.
- Licence
- GNU Lesser General Public License 3.0, which builds on the GNU General Public License 3.0
- Source
- github.com/zhuker/lamejs
It is kept unchanged and separate in the audio editor's page, as its licence asks.
-
gifenc 1.0.3
Saves animated GIFs in the video editor.
- Licence
- MIT License
- Source
- github.com/mattdesl/gifenc
- Copyright
- Matt DesLauriers
Models That Download When You Use Them
"Write subtitles for me" in the video editor, Transcribe in the audio editor and Background in the photo editor use these. They download only after you agree, the first time you use the feature: the code from jsDelivr, the models from Hugging Face. They run on your device. Your video, your audio and your photos are never sent anywhere.
-
Whisper base.en and Whisper base
The speech recognition models that write out what is said: base.en for English, base (multilingual) for the other languages and Auto-detect. Each is an 8-bit ONNX copy of about 80 MB.
- Licence
- MIT License
- Source
- github.com/openai/whisper, model files from huggingface.co/onnx-community/whisper-base.en (revision 51eefc0) and huggingface.co/onnx-community/whisper-base (revision 1846881)
- Copyright
- OpenAI
-
MODNet
The model that finds the person in a photo for Background: an ONNX copy of about 25 MB, at a fixed version (revision fa2fa54).
- Licence
- Apache License 2.0
- Source
- github.com/ZHKKKe/MODNet, model file from huggingface.co/Xenova/modnet
- Copyright
- Zhanghan Ke and contributors
-
Transformers.js 4.3.1
Runs the speech model in your browser, with ONNX Runtime Web.
- Licence
- Apache License 2.0
- Source
- github.com/huggingface/transformers.js
- Copyright
- Hugging Face
-
ONNX Runtime Web
The engine that runs both models in your browser: version 1.30.0 for Background, and for subtitles the version Transformers.js 4.3.1 brings along.
- Licence
- MIT License
- Source
- github.com/microsoft/onnxruntime
- Copyright
- Microsoft Corporation
Fonts
All six fonts are under the SIL Open Font License 1.1, which allows using and embedding them for free. The files are used unchanged and come from this site. Letters they don't have, such as Arabic, Hindi, Bengali, Thai, Chinese, Japanese and Korean, use your device's own fonts; nothing is downloaded for them.
-
Inter 4.001
The text of the site and the editors.
- Licence
- SIL Open Font License 1.1
- Source
- github.com/rsms/inter
-
Anton 2.116
A subtitle and end card font.
-
Archivo Black 1.006
A subtitle and end card font.
-
Merriweather 2.100
A subtitle and end card font. Merriweather is a Reserved Font Name, which is one more reason the files stay unchanged.
-
Oswald 4.103
Supplies the Cyrillic letters for Anton, which has none.
-
Montserrat 9.000
Supplies the Cyrillic letters for Archivo Black, which has none.
Full Licence Texts
- Mozilla Public License 2.0
- GNU Lesser General Public License 3.0
- GNU General Public License 3.0
- Apache License 2.0
- MIT License, one copy per part with its copyright line: gifenc, Whisper, ONNX Runtime Web
- SIL Open Font License 1.1, one copy per font with its copyright line: Inter, Anton, Archivo Black, Merriweather, Oswald, Montserrat
The Mozilla text comes from mozilla.org and the Apache text from apache.org. The two GNU texts are the SPDX licence list's copies of the official texts (github.com/spdx/license-list-data).