Web Playground
The Daemo Playground is a visual chat interface at app.daemo.ai where you can test your agent without writing any code.
- Quick demos to stakeholders
- Non-technical team members testing
- Exploring agent capabilities
- Sharing test results
- Rapid development iteration
- Automated testing
- Scripted test suites
- CI/CD pipelines
Accessing the Playground
- Go to app.daemo.ai
- Select your Organization
- Click on your Agent
- Click Playground in the sidebar (marked with a
TESTINGbadge)
Your service must be running and connected (showing 🟢 Online status) for the Playground to work.
Using the Playground
The Query Playground Interface
When you open the Playground, you'll see a clean chat interface. Your agent greets you automatically:
Query Playground
Test your agent with custom configurations
Basic Conversation
Type natural language queries and watch the agent respond:
I found 3 contacts at Acme Corp:
• John Smith - john@acme.com (CEO)
• Sarah Johnson - sarah@acme.com (CTO)
• Mike Wilson - mike@acme.com (Sales)
Multi-Turn Conversations
The Playground maintains conversation context:
Viewing Function Calls
The Playground shows which functions are called:
{ start_date: '2024-01-01', end_date: '2024-01-31' }{ deals: [...], count: 5 }Configuration Panel
Click the ⚙️ Configure button in the top-right corner to open the Configuration panel. This lets you customize how your agent responds to queries.
Configuration Options
| Option | Description |
|---|---|
| Agent Role | Set a role context (e.g., "support", "analyst", "developer") to influence agent behavior |
| Max Tokens | Limit the maximum response length (default: 4096) |
| Custom LLM | Use a custom LLM configuration instead of backend defaults |
| Query Context (JSON) | Pass additional context data as JSON that your functions can access |
| Analysis Mode | ✨ Enable deeper analysis and reasoning in responses |
| Use Streaming | ✨ Stream responses in real-time as they're generated |
Query Context is useful for passing user-specific data like user_id, org_id, or session information that your functions need.
Free Playground Testing
💡 Playground queries are free and don't deduct credits. Use the playground as much as you need during development and testing without worrying about usage costs.
Playground Features
Clear Conversation
Click the Clear button to start fresh. This resets context.
Export Conversation
Copy or export conversations for documentation or bug reports.
Switch Services
If you have multiple services connected, switch between them to test different capabilities.
Service Status Indicators
| Status | Meaning |
|---|---|
| 🟢 Online | Service connected and ready |
| 🔴 Offline | Service not running |
| 🟡 Connecting | Establishing connection |
If your service shows Offline, the Playground won't work. Start your service first:
npm run dev
# or
npx ts-node src/index.ts
Playground vs CLI Testing
| Feature | Playground | CLI |
|---|---|---|
| Setup required | None | Script + tools |
| Iteration speed | Medium | Fast |
| Shareable | Yes (screenshots) | No |
| Scriptable | No | Yes |
| See raw events | Limited | Full stream |
| Best for | Demos, exploration | Development |
Recommended workflow: Use CLI during active development, Playground for demos and verification.
Tips for Effective Playground Testing
1. Start Simple
Begin with straightforward queries before testing complex scenarios:
✓ "List all contacts"
✓ "Find contact john@example.com"
✓ "How many deals do we have?"
2. Test Edge Cases
Try queries that might break things:
"Find contact with ID invalid-uuid"
"Delete all contacts" (should your agent allow this?)
"What's the weather?" (outside your agent's scope)
3. Verify Function Selection
Watch the function calls panel to ensure the right functions are being called.