SnuhiLearn is built on open-source software and an open-weight AI model, used under their licences. The principal components:
Qwen2.5-14B-Instruct-AWQ (Alibaba Cloud / Qwen) - AI language model, self-hosted - Apache License 2.0
vLLM - Model serving runtime - Apache License 2.0
Python 3 - Server language runtime - PSF License
FastAPI - Web framework - MIT
Starlette - ASGI toolkit - BSD-3-Clause
Uvicorn - ASGI server - BSD-3-Clause
Pydantic - Data validation - MIT
SQLAlchemy - Database layer - MIT
httpx - HTTP client - BSD-3-Clause
python-multipart - Form parsing - Apache License 2.0
pypdf - PDF text extraction - BSD-3-Clause
SQLite - Database engine - Public domain
Tesseract OCR - Photo-doubt text reading - Apache License 2.0
Flutter & Dart SDK - Mobile app framework - BSD-3-Clause
http / shared_preferences / image_picker / flutter_secure_storage (Flutter plugins) - Mobile app - BSD-3-Clause
provider / file_picker / audioplayers / cupertino_icons (Flutter plugins) - Mobile app - MIT
The tutor's language model is Qwen2.5-14B-Instruct-AWQ, licensed by the Qwen team (Alibaba Cloud) under the Apache License 2.0. Under that licence the model may be used commercially, including for a paid service, without royalties; this notice records the attribution the licence requires. The full licence text is available at https://www.apache.org/licenses/LICENSE-2.0 and ships with the model.
External services used under their own terms (not open source): Brevo (email delivery), Meta WhatsApp Cloud API (messaging, when enabled by the operator), Let's Encrypt (TLS certificates). No copyleft-licensed (GPL/AGPL) component is shipped or linked by this product. The canonical inventory is maintained as NOTICE.md in the product's source repository.