AgentVista Documentation

Unified observability for AI agents and infrastructure — one platform to monitor, trace, and correlate everything.

AgentVista gives you a single place to observe your AI agents and the infrastructure beneath them. Stop context-switching between LangSmith for agent traces and Grafana for infrastructure metrics. With AgentVista, one SDK call instruments your agents, and one OTLP endpoint receives everything else.

Quick Start

Send your first trace in under 5 minutes.

Authentication

Create an API key and start sending data.

SDK Reference

Install the Python SDK and explore all tracing APIs.

API Reference

Explore REST endpoints for ingestion and the dashboard.

Get started in 3 steps

1. Install the SDK

   pip install agentvista
   
2. Initialize with your API key

   import agentvista
   agentvista.init(api_key="av_xxxxx")
   

   Get your API key from Settings → API Keys in the AgentVista dashboard.

3. Instrument your first agent

   @agentvista.trace_agent
   def my_agent(prompt):
       response = llm_call(prompt)
       return response
   

   Every call to my_agent now appears as a trace in your dashboard.

Explore by topic

Tracing

Use decorators, context managers, and child spans to model your agent workflows.

Adapters

Auto-extract token counts and cost from Anthropic, OpenAI, and LangChain responses.

OpenTelemetry

Point any OTel-instrumented service at AgentVista using standard OTLP.

Alerts

Set threshold and composite alerts on cost, failure rate, and latency.

Prompt Versioning

Track which prompt version produced which outcomes and compare side-by-side.

Service Map

Auto-discover your service topology — AI agents alongside databases and APIs.