zynex-ai: Skill for Claude Code

Testing Zynex AI

Prerequisites

• Python 3.x with dependencies from requirements.txt installed
• No .env file needed for demo mode testing (app auto-detects and runs in demo mode)

Devin Secrets Needed

• OPENROUTER_API_KEY (optional): Only needed to test the full OAuth key exchange flow end-to-end. Without it, you can still verify demo mode, banner UI, popup trigger, and model switching.

Starting the Server

cd /home/ubuntu/repos/zynex-ai
pip install -r requirements.txt
python run.py
# Server runs on http://localhost:8000

If port 8000 is already in use, kill the existing process first:

lsof -ti:8000 | xargs kill -9 2>/dev/null
python run.py

Key Test Areas

1. Demo Mode (No API Key)

• Navigate to http://localhost:8000
• Verify the "Activate Free AI" banner appears at the top with a "FREE" badge
• Verify /api/config returns {"demo_mode":true,"has_api_key":false,"default_model":"google/gemini-2.0-flash-001"}
• Send a chat message — should receive a structured demo response with sections, bullet points
• Demo response should end with italic text: "Running in demo mode — activate free AI via the setup banner for real responses."

2. OAuth Flow

• Click "Activate Free AI" button
• A popup should open to openrouter.ai/auth (or openrouter.ai/sign-up with redirect)
• The callback URL in the popup URL should be URL-encoded (look for %253A, %252F)
• The button should change to "Connecting…" (disabled) while the popup is open
• After closing the popup without authenticating, the button should revert to "Activate Free AI"
• Full OAuth test (requires OpenRouter account): Complete the auth flow, verify key is saved to .env, config hot-reloads, and the banner disappears

3. Model Switcher

• Click the model selector pill in the header (e.g., "Gemini 2.0 (Free)")
• It should cycle through: Gemini 2.0 → Llama 3.3 → DeepSeek R1 → Gemini 2.0
• A toast notification should appear for each switch (e.g., "Switched to Llama 3.3 (Free)")