pyr0ball
cd2fdcc45b
- Add MIT LICENSE (entire current codebase is MIT-safe per cf-licensing review) - Add THIRD_PARTY_LICENSES.md (Precise/Whisper/Piper/pyannote/MaixPy) - Add CLAUDE.md to .gitignore (BSL 1.1 + docs-location policy) - Fix maix_voice_client.py: load credentials from secrets.py instead of inline constants (matches existing secrets.py.example pattern) - Remove CLAUDE.md from git tracking
36 lines
1.4 KiB
Markdown
36 lines
1.4 KiB
Markdown
# Third-Party Licenses
|
|
|
|
## Mycroft Precise
|
|
|
|
- **License:** Apache 2.0
|
|
- **Source:** https://github.com/MycroftAI/mycroft-precise
|
|
- **Usage:** Wake word detection engine. Downloaded at setup time; not bundled in this repo.
|
|
- **Note:** If you distribute a packaged Minerva build that includes the `precise-engine` binary, you must include the Apache 2.0 NOTICE file from the Mycroft Precise repository.
|
|
|
|
## OpenAI Whisper
|
|
|
|
- **License:** MIT
|
|
- **Source:** https://github.com/openai/whisper
|
|
- **Usage:** Speech-to-text inference.
|
|
|
|
## Piper TTS
|
|
|
|
- **License:** MIT
|
|
- **Source:** https://github.com/rhasspy/piper
|
|
- **Usage:** Text-to-speech synthesis.
|
|
|
|
## pyannote.audio
|
|
|
|
- **License:** MIT (model weights require separate HuggingFace user agreement)
|
|
- **Source:** https://github.com/pyannote/pyannote-audio
|
|
- **Usage:** Speaker identification and diarization (optional, in `voice_server_enhanced.py`).
|
|
- **Note:** To use speaker identification you must accept pyannote's terms on HuggingFace
|
|
and provide your own `HF_TOKEN`. Minerva does not distribute pyannote model weights.
|
|
|
|
## MaixPy / K210 Firmware
|
|
|
|
- **License:** Apache 2.0 (MaixPy); some Kendryte/Sipeed components are proprietary.
|
|
- **Source:** https://github.com/sipeed/MaixPy
|
|
- **Usage:** MicroPython firmware for Maix Duino (K210) edge devices.
|
|
- **Note:** Firmware is NOT included in this repository. Obtain directly from Sipeed:
|
|
https://dl.sipeed.com/MAIX/MaixPy/release/
|