Dashboard Features
Traces
View detailed execution traces showing the complete flow of your agent’s logic. Each trace includes:- Input/output for every step
- Token usage and latency
- Error states and stack traces
- Parent-child span relationships
Sessions & Conversations
Group related interactions together to understand user journeys and debug multi-turn conversations.Analytics
Track key metrics over time:- Total requests and error rates
- Token usage by model
- Average latency
- Cost per agent
How It Works
The AgentBasis SDK uses OpenTelemetry (OTel) under the hood, the industry standard for observability.- Instrument your code with a single function call per LLM provider or framework
- Data is batched and sent asynchronously (minimal latency impact on your app)
- View traces in the AgentBasis dashboard in real-time
Use Cases
Production Monitoring
Track agent performance, error rates, and costs in real-time
Debugging
Trace issues back to specific users, sessions, or inputs
Cost Optimization
Identify expensive prompts and optimize token usage
Compliance
Audit logs of all AI interactions with user context
Next Steps
Quick Start
Install the SDK and start tracking in 5 minutes
API Reference
Complete list of all SDK functions

