Built on good shoulders.
Otterlog is proprietary software, but it stands on dozens of open-source libraries, models, and tools. They're listed in full below, with their licenses and attribution. Every component has been audited for commercial-use compatibility, and we don't use anything with a copyleft (GPL/LGPL/AGPL) licence.
Last reviewed: April 2026. Synced with
THIRD_PARTY_LICENSES.md in the Otterlog source tree.
Application core.
| Component | License | Copyright |
|---|---|---|
| Tauri v2 | MIT or Apache-2.0 | Tauri Contributors |
| whisper-rs | Unlicense | Tim Huff |
| whisper.cpp | MIT | Georgi Gerganov |
| sherpa-onnx | Apache-2.0 | k2-fsa / Next-gen Kaldi |
| llama-cpp-2 | MIT or Apache-2.0 | llama-cpp-rs contributors |
| llama.cpp | MIT | Georgi Gerganov |
| cpal | Apache-2.0 | RustAudio Developers |
| hound | MIT or Apache-2.0 | Ruud van Asseldonk |
| rubato | MIT | rubato contributors |
| rusqlite | MIT | rusqlite developers |
| parking_lot | MIT or Apache-2.0 | Amanieu d'Antras |
| serde | MIT or Apache-2.0 | Serde Developers |
| tokio | MIT | Tokio Contributors |
| anyhow | MIT or Apache-2.0 | David Tolnay |
| chrono | MIT or Apache-2.0 | chrono developers |
| uuid | MIT or Apache-2.0 | uuid-rs contributors |
| clap | MIT or Apache-2.0 | clap contributors |
User interface.
| Component | License | Copyright |
|---|---|---|
| React | MIT | Meta Platforms, Inc. |
| Tailwind CSS | MIT | Tailwind Labs, Inc. |
| Lucide | ISC | Lucide contributors |
| Vite | MIT | Evan You |
| TypeScript | Apache-2.0 | Microsoft Corporation |
Speaker separation.
| Component | License | Copyright |
|---|---|---|
| pyannote.audio v4.x | MIT | CNRS / pyannote.ai |
| PyTorch | BSD-3-Clause | Meta Platforms, Inc. |
| SpeechBrain | Apache-2.0 | SpeechBrain contributors |
| soundfile | BSD-3-Clause | Bastian Bechtold |
The intelligence layer.
| Model | License | Copyright |
|---|---|---|
| OpenAI Whisper | MIT | OpenAI |
| Silero VAD v5 | MIT | Silero Team |
| pyannote/speaker-diarization-3.1 | MIT | CNRS / pyannote.ai |
| pyannote/speaker-diarization-community-1 | CC-BY-4.0 | pyannote.ai |
| pyannote-segmentation-3.0 | MIT | CNRS / pyannote.ai |
| wespeaker VoxCeleb ResNet34-LM | CC-BY-4.0 | WeNet/WeSpeaker contributors |
| Qwen2.5-7B-Instruct | Apache-2.0 | Alibaba Cloud |
Required notice.
Two of the models above are used under the Creative Commons Attribution 4.0 International License. That licence requires us to provide visible attribution wherever Otterlog is distributed; this is that attribution.
pyannote/speaker-diarization-community-1
- Created by: Hervé Bredin and pyannote.ai contributors
- Source: huggingface.co/pyannote/speaker-diarization-community-1
- License: CC-BY-4.0
- Usage in Otterlog: Speaker diarization PLDA model weights, used unmodified.
wespeaker VoxCeleb ResNet34-LM Embedding Model
- Created by: WeNet/WeSpeaker contributors
- Source: github.com/wenet-e2e/wespeaker
- License: CC-BY-4.0
- Usage in Otterlog: Speaker embedding extraction for diarization, used unmodified.
How we stay compliant.
- Attribution preserved. Full license texts are bundled with the application and the source tree.
- No copyleft. Otterlog uses no GPL/LGPL/AGPL components. Nothing in the dependency tree requires us to open-source Otterlog itself.
- Commercial compatibility verified. Every component above permits commercial use. The earlier Qwen2.5-3B-Instruct model (research-only "Tongyi Qianwen" license) was identified as non-commercial during audit and replaced with Qwen2.5-7B-Instruct (Apache-2.0).
- Gated HuggingFace models. Where models are gated, the Papi Labs developer has accepted the model licence on the user's behalf. End users don't need HuggingFace accounts.
Found a missing attribution or licence error? Email hello@otterlog.ai and we'll correct it within 7 days.