TechFixAI is a production-grade incident management platform that bridges the language gap in Japanese software teams. It captures voice-recorded incident reports in Japanese, transcribes and translates them using Groq Whisper with technical context awareness, and automatically generates structured support tickets — all in one seamless background pipeline.

The platform goes beyond simple transcription. After a voice report is submitted, TechFixAI deterministically assigns the right developer based on incident type and severity — no random shuffling, no manual triage. Built on FastAPI with SQLAlchemy and Jinja2-rendered pages, it supports SQLite locally and PostgreSQL in production.

TechFixAI ships production-ready: full auth with Google OAuth, rate limiting, CAPTCHA, security headers, structured logging, Sentry integration, and Prometheus metrics — all configured and ready to deploy.

• ✅ Voice upload pipeline — Audio → STT → Translation → Ticket → Assignment as a background task
• ✅ Japanese-to-English translation with technical context awareness
• ✅ Deterministic developer assignment by incident type and severity
• ✅ Chat voice helper — transcribe and translate without creating a ticket
• ✅ Full auth system — email/password + Google OAuth + secure session cookies
• ✅ CAPTCHA support on login and signup (Cloudflare Turnstile style)
• ✅ Granular rate limiting for global traffic, auth, uploads, and translation
• ✅ Admin dashboard with aggregated stats, conversations, and developer metrics
• ✅ Health and Prometheus metrics endpoints
• ✅ Structured logging with request IDs and sensitive value redaction
• ✅ Security headers, HTTPS enforcement, CORS validation, brute-force lockout
• ✅ Optional Sentry integration for error tracking and tracing

• FastAPI — Core web framework and API routing
• SQLAlchemy — ORM and database abstraction
• Starlette — Middleware (request ID, security headers, rate limiting)
• Jinja2 — Server-rendered HTML templates
• Uvicorn — ASGI server

• SQLite — Local development default
• PostgreSQL — Production via DATABASE_URL

• Groq Whisper — Primary speech-to-text transcription
• OpenAI — Optional fallback
• Gemini — Optional fallback

• Render — Deployment target
• Google OAuth 2.0 — Social login
• Sentry — Optional error tracking and tracing
• Prometheus — Optional metrics scraping

TechFixAI/
├── app/
│   ├── main.py                  # App entry point, route registration
│   ├── models/                  # SQLAlchemy models
│   ├── routes/
│   │   ├── auth.py              # Login, signup, logout, Google OAuth
│   │   ├── voice.py             # Voice upload, chat-transcribe, status
│   │   ├── tickets.py           # Ticket CRUD and status management
│   │   ├── admin.py             # Dashboard, conversations, developers
│   │   └── pages.py             # Jinja2 page routes
│   ├── services/
│   │   ├── stt.py               # Groq Whisper STT integration
│   │   ├── translation.py       # Technical-context translation
│   │   ├── ticket_generator.py  # Ticket structuring logic
│   │   └── assignment.py        # Deterministic developer assignment
│   ├── middleware/
│   │   ├── rate_limit.py        # Route-specific rate limiters
│   │   ├── request_id.py        # Request ID injection
│   │   └── security.py          # Security headers, HTTPS enforcement
│   ├── templates/               # Jinja2 HTML templates
│   │   ├── login.html
│   │   ├── signup.html
│   │   ├── upload.html
│   │   ├── dashboard.html
│   │   ├── tickets.html
│   │   └── developers.html
│   └── config.py                # Settings and env variable validation
├── .github/
│   └── workflows/
│       ├── ci.yml
│       ├── secret-hygiene.yml
│       └── uptime-checks.yml
├── requirements.txt
└── README.md

# Clone the repository
git clone https://github.com/LegendarySumit/TechFixAI.git
cd TechFixAI

# Create and activate virtual environment
python -m venv .venv

# Windows
.venv\Scripts\activate
# macOS/Linux
source .venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Run the development server
uvicorn app.main:app --reload

Open → http://127.0.0.1:8000

SECRET_KEY=your_secret_key
DATABASE_URL=postgresql://...
CORS_ORIGINS=https://your-frontend.vercel.app

PUBLIC_DEPLOYMENT=true
FORCE_HTTPS=true
SESSION_COOKIE_SECURE=true
APP_BASE_URL=https://your-backend-domain.onrender.com
ADMIN_EMAILS=[email protected],[email protected]

GROQ_API_KEY=your_groq_key        # Required
OPENAI_API_KEY=your_openai_key    # Optional fallback
GEMINI_API_KEY=your_gemini_key    # Optional fallback

GOOGLE_CLIENT_ID=your_client_id
GOOGLE_CLIENT_SECRET=your_client_secret
GOOGLE_REDIRECT_URI=https://your-backend-domain.onrender.com/auth/google/callback

SMTP_HOST=smtp.example.com
SMTP_PORT=587
SMTP_USER=[email protected]
SMTP_PASSWORD=your_password
FROM_EMAIL=[email protected]

⚠️ APP_BASE_URL must point to the backend URL. CORS_ORIGINS must contain frontend origin(s). No trailing slashes.

1. Create a Neon Postgres project and copy the pooled connection string.
2. In Render, create a new Web Service from this repository.
3. Keep these commands:
   • Build: pip install -r requirements.txt
   • Start: uvicorn app.main:app --host 0.0.0.0 --port $PORT --proxy-headers --forwarded-allow-ips=*
4. Set DATABASE_URL in Render using your Neon connection string.
5. Set required env vars in Render: SECRET_KEY, APP_BASE_URL, CORS_ORIGINS, GROQ_API_KEY (plus OAuth/SMTP if used).
6. Set health check path to /health.

Example Neon URL format:

DATABASE_URL=postgresql://<user>:<password>@<neon-host>/<db>?sslmode=require

This repository now includes a Render blueprint at render.yaml for backend-only deployment.

1. Sign up or log in at /signup or /login — Google OAuth available
2. Upload a voice report at /upload — select a Japanese audio file and submit
3. The pipeline runs in the background: transcription → translation → ticket → assignment
4. Track tickets at /tickets — click any ticket for full detail
5. Chat transcription — use the voice helper to transcribe without creating a ticket
6. Admin dashboard at /dashboard — stats, developer workload, conversation logs

Assignment is fully rule-based — no randomness:

1. Backend/infrastructure + high or critical → Backend team
2. Frontend/UI incident → Frontend team
3. Database + high or critical → Backend team
4. Fallback → Least loaded developer

Startup fails during settings load?

• Verify CORS_ORIGINS is not empty and uses a valid format
• Ensure APP_BASE_URL points to the backend domain, not the frontend
• Confirm DATABASE_URL is a valid PostgreSQL connection string

Voice upload not processing?

• Check that GROQ_API_KEY is set and valid
• Poll /api/voice/status/{conversation_id} for pipeline progress

Google OAuth failing?

• Ensure GOOGLE_REDIRECT_URI matches exactly what's registered in Google Cloud Console

• Multi-language support beyond Japanese
• Slack / Teams notifications on ticket creation
• SLA tracking and breach alerts
• ML-based developer assignment
• Mobile app for on-the-go voice reporting
• Webhook support for external integrations
• Role-based access control (RBAC) for teams

This project is licensed under a Proprietary All Rights Reserved License. See the LICENSE file for details.

LegendarySumit

• GitHub: @LegendarySumit
• Project: TechFixAI
• Live Demo: https://techfixai.onrender.com/