Leverage z.ai team of experts for domain-specific questions about trading, finance, engineering, and security
z.ai provides a team of domain experts accessible through the zread MCP server, working alongside Google Antigravity IDE.
Status: Configured and ready to use
# Session only
$env:Z_AI_API_KEY = "your-z-ai-key"
# Permanent (User-level)
[Environment]::SetEnvironmentVariable("Z_AI_API_KEY", "your-key", "User")
Located in .antigravity/mcp-config.json:
{
"zread": {
"type": "streamableHttp",
"url": "https://api.z.ai/api/mcp/zread/mcp",
"headers": {
"Authorization": "${Z_AI_API_KEY}"
}
}
}
| Expert | Domain | Best For |
|---|---|---|
| 💰 Finance | Markets, Trading | Strategy, risk management, orderflow patterns |
| ⚙️ Engineering | Architecture, Code | System design, performance, debugging |
| 📊 Data Science | ML, Analytics | Signal processing, pattern detection, models |
| 🔐 Security | Compliance, Safety | API security, secrets, audit logging |
| Scenario | Expert | Example Query |
|---|---|---|
| Trading logic | Finance | "What orderflow patterns indicate institutional accumulation?" |
| System design | Engineering | "How should I structure WebSocket reconnection for HFT?" |
| ML features | Data Science | "What features best predict short-term price movements?" |
| Security review | Security | "What's the best approach for rotating API keys in production?" |
graph LR
A[Question] --> B{Type?}
B -->|Trading/Market| C[z.ai Finance]
B -->|Code/Architecture| D[z.ai Engineering]
B -->|ML/Data| E[z.ai Data Science]
B -->|Security/Compliance| F[z.ai Security]
C --> G[Expert Analysis]
D --> G
E --> G
F --> G
G --> H[Antigravity Implements]
"I'm building an orderflow system for Indian equities. What footprint
patterns should I detect to identify institutional activity, given
the 5% circuit limit and market hours of NSE?"
"My WebSocket connection to Alpaca drops every few hours. What's the
best reconnection strategy for a trading system where I can't miss
more than 1 second of L2 data?"
[!CAUTION]
- z.ai provides expertise, not financial advice
- Always validate recommendations against your requirements
- Test thoroughly in paper trading before live deployment
- Expert input should inform, not replace, your judgment
