Licences & third-party notices
Last updated 23 October 2026
TinyPine Studio
© PineCore Systems, a trading name of Bröderna Saxin AB. Free to use under the Terms of Service.
Models
No model ships inside the installer. You choose what to install in the Grove, and each model's licence file is installed next to its weights in your library.
| Model | Used for | Licence |
|---|---|---|
| Krea 2 Turbo, Krea 2 Raw (Krea) | Image generation, editing, LoRA training | Krea 2 Community License Agreement |
| LTX-2.5 (Lightricks) with its text encoder, VAEs and upscalers | Video with sound | LTX-2.x Community License Agreement |
| MiniMax-Music3 (MiniMax) | Music | MiniMax-Music3 Community License |
| AudioSeal (Meta) | Watermark in every song | MIT |
| Qwen3-VL 4B and 8B Instruct, abliterated (huihui-ai, from Qwen) | TinyPine Assist | Apache-2.0 |
| Depth Anything V2 Small | Depth for bokeh, fog and relight | Apache-2.0 |
| GFPGAN v1.4 with RetinaFace (facexlib) | Face Enhance | Apache-2.0 / MIT |
| 2xNomosUni SPAN (Phips) | Upscaler | See the model card |
| RelayVSR 2x: native 2x checkpoints by LightX2V, method by Xijun Wang et al., on a Wan 2.2 5B base | Upscale video, and the ×2 step after a video | RelayVSR code and Wan 2.2: Apache-2.0. The LightX2V weights: no licence stated on the repository |
| RIFE 4.25 (hzwer, Practical-RIFE) | Slow motion, Fast | MIT |
| FILM (Google Research), PyTorch export by dajes | Slow motion, Quality | Apache-2.0 |
| Ming Image 0.1 Design and Design-Layer (inclusionAI) | Design workspace, split into layers | MIT |
| SAM 2.1 Hiera Base Plus (Meta) | Smart select in Design, masks, moving masks on video | Apache-2.0 |
| Lucida v6, a BiRefNet fine-tune | Cut out, subject and background | MIT |
| MediaPipe Face Landmarker (Google) | Face in Edit | Apache-2.0 |
| BootsTAPIR (Google DeepMind, TAPNet) | Tracking | Apache-2.0 |
| MOSS-TTS v1.5, MOSS-VoiceGenerator and the MOSS audio tokenizers (OpenMOSS) | Speech in 31 languages and voice design | Apache-2.0 |
| Chatterbox (Resemble AI) | English speech | MIT; each clip also carries Resemble's Perth watermark |
| EzAudio-XL with its audio codec | Sound effects | MIT / Apache-2.0, see the model card |
| Lensfun lens database | Lens profiles in Edit (distortion, vignetting, colour fringes) | CC BY-SA 3.0, shipped unmodified |
| Whisper large-v3-turbo and small (OpenAI), KB-Whisper (KBLab) | Transcription and subtitles | MIT / Apache-2.0, see the model cards |
Downloads come from PineCore Systems mirrors on Hugging Face, which are unmodified copies carrying the original licence and notice files, with the original repositories as fallback. The RelayVSR 2x weights come directly from LightX2V's repository (huggingface.co/lightx2v/RelayVSR_X2), with a notice file naming the sources written next to them.
Krea 2 notice
Krea 2 is licensed under the Krea 2 Community License Agreement. For more information, visit https://krea.ai/krea-2-licensing.
The Krea 2 licence covers organisations with up to USD 1,000,000 in total revenue over the trailing twelve months. The LTX-2.x licence covers organisations with up to USD 10,000,000 in annual revenue. The MiniMax-Music3 licence requires a separate authorisation from MiniMax above USD 20,000,000 in yearly revenue. See the Acceptable Use Policy for what they prohibit.
MiniMax-Music3 notice
Music is made with MiniMax-Music3, Copyright (c) 2026 MiniMax, used under the MiniMax-Music3 Community License and its Acceptable Use Policy (Exhibit A). The licence text is installed next to the weights.
Open-source components
| Component | Role | Licence |
|---|---|---|
| GPUI (Zed Industries), GPUI Component and GPUI Kit (Longbridge) | The app's interface | Apache-2.0 |
| Rust crates, among them tiny-skia, rustybuzz, ab_glyph, resvg, glow, cpal, png, webp | Drawing, text shaping, design files, sound output | MIT / Apache-2.0 / BSD-3-Clause / Zlib, see each crate |
| krilla with pdf-writer, subsetter, skrifa, read-fonts, write-fonts and xmp-writer | Vector PDF from Design | MIT / Apache-2.0 |
| tiny-skia-path | Shapes as paths in vector PDF | BSD-3-Clause |
| clap-sys, vst3 and com-scrape-types | Hosting CLAP and VST3 plugins | MIT / Apache-2.0 |
| VST 3 API (Steinberg Media Technologies GmbH) | Hosting VST3 plugins | MIT |
| libloading | Loading plugins | ISC |
| midir | MIDI input | MIT |
| ANGLE (Google) | OpenGL ES for the effect engine | BSD-3-Clause |
| Next.js, React | This website | MIT |
| Python (embedded) | Engine runtime | PSF License |
| PyTorch, torchvision, torchaudio | GPU inference | BSD-3-Clause |
| diffusers, transformers, accelerate, safetensors, huggingface_hub, peft | Model loading, inference, LoRA | Apache-2.0 |
| torchao | fp8 precision | BSD-3-Clause |
| bitsandbytes | 4-bit Assist and training | MIT |
| LTX-2 (ltx-core, ltx-pipelines) | Video pipeline | See the Lightricks/LTX-2 repository |
| RelayVSR inference code (kopperx/RelayVSR), shipped with the app | Upscale video | Apache-2.0 |
| Triton (triton-windows), einops, PyAV | Upscale video: GPU kernels, tensors, reading and writing video | MIT / MIT / BSD-3-Clause |
| audioseal (Meta) | Watermark in every song | MIT |
| soundfile, libsndfile | Writing song files | BSD-3-Clause / LGPL-2.1 |
| Wikidata (artist and song names) | The music safeguards' name lists, shipped with the app | CC0 1.0 |
| spandrel, facexlib | Upscaler and face models | MIT |
| OpenCV | Image processing (via facexlib) | Apache-2.0 |
| FFmpeg (via imageio-ffmpeg) | Video encoding and export | LGPL / GPL, depending on the build |
| FastAPI, Uvicorn, Pydantic | Local engine server | MIT / BSD |
| Pillow | Image files | MIT-CMU (HPND) |
| cryptography | Signature checks for the Grove | Apache-2.0 / BSD |
| faster-whisper, CTranslate2 | Transcription | MIT |
| chatterbox-tts, resemble-perth, s3tokenizer, conformer, librosa | English speech (Chatterbox runtime) | MIT / Apache-2.0 / ISC |
| OpenCV (opencv-python) | Photo Merge: HDR and panorama | Apache-2.0 |
| MediaPipe | Face landmarks | Apache-2.0 |
| httpx, pypdf, python-docx, python-multipart | Local file and network helpers | BSD / MIT / Apache-2.0 |
| Phosphor Icons | Icons | MIT |
| Bricolage Grotesque, Hanken Grotesk, JetBrains Mono, and the 26 typefaces bundled with Design | Typefaces | SIL Open Font License 1.1 |
Each component stays under its own licence. This page is a summary and does not replace them.
VST is a registered trademark of Steinberg Media Technologies GmbH. CLAP is an open plugin standard by Bitwig and u-he. TinyPine hosts plugins you install yourself and does not include any.