ADK Django

python3 -m venv .venv
source .venv/bin/activate       # ubuntu
.\.venv\Scripts\activate        # powershell

pip install -r requirements.txt
python3 manage.py makemigrations
python3 manage.py migrate

python3 manage.py collectstatic
python3 manage.py createsuperuser
daphne myadk.asgi:application -b 0.0.0.0 -p 8000

Surface Web

Dark Web

Flutter App

To do List:

1. Dynamic UI & BYOK Data Model (Django): Configure a single TenantCompany model to store all dynamic branding (logo URL, colors) and per-tenant settings (LLM choice, System Prompt). Authenticated Django views must fetch and pass this model's data via the context dictionary for all UI rendering and backend API calls (Bring Your Own Key - BYOK).

2. Voice AI Orchestration (Flask/Webhooks): Implement a separate, lightweight Flask application exposed via a public URL (e.g., Twilio webhook). This app must contain two endpoints, start_call and process_prompt, that use TwiML/NCL to manage the call flow, perform Speech-to-Text (STT), call the Groq/Gemini API (using the key from the Django DB), and perform Text-to-Speech (TTS).

3. Cross-Platform Data Integration: Ensure the Flask (call) logic accesses the same Django database (e.g., using PostgreSQL/MySQL container setup instead of SQLite) to retrieve the user's specific API key and LLM model name for reply generation, unifying the settings across both the chat and call interfaces.