Creditondairy/trainer

title: Factory Voice Trainer emoji: 🎧 colorFrom: indigo colorTo: blue sdk: streamlit sdkversion: "1.39.0" appfile: app.py pinned: false ---

Factory Voice Trainer (External Version - Starter)

This is a minimal Streamlit app that lets operators chat with a Factory Machine Voice Trainer. It supports:

• —Uploading SOPs (PDF/TXT)
• —Context-grounded guidance (starter keyword retriever)
• —Optional voice replies via OpenAI TTS

For production, swap the simple retriever for embeddings (OpenAI + Chroma/FAISS) and add microphone input (ASR) with streamlit-webrtc + Whisper.

🔧 Local Run

1) Python 3.10+ recommended. 2) Install deps:

pip install -r requirements.txt

3) Create .env with your OpenAI key:

OPENAI_API_KEY=sk-...

4) Run:

streamlit run app.py

☁️ Deploy (Hugging Face Spaces - no GitHub required)

1) Create a new Space (type: Streamlit). 2) Upload these files: app.py, requirements.txt, README.md. 3) In Settings → Secrets, add your key:

OPENAI_API_KEY=sk-...

4) The app builds and provides a public URL.

🗺️ Roadmap

• —Replace keyword retriever with embeddings (OpenAI text-embedding-3-large) + vector DB.
• —Add ASR (live microphone) with streamlit-webrtc + Whisper.
• —Add operator profiles + progress logging (Firestore/SQL).
• —Build step engine for Start-up / CIP / Shutdown with checkpoints.