Agno - Zespan

Available via: OpenTelemetry auto-instrumentation (Python). This is not a Zespan SDK wrapper — see Send traces from any OTel-instrumented app for how the underlying endpoint works.

Agno is a fast-growing multi-agent framework. OpenInference’s openinference-instrumentation-agno package instruments agent runs, team coordination, and tool calls — point the exporter at Zespan and every run traces automatically.

Installation

pip install agno openinference-instrumentation-agno \
  opentelemetry-sdk opentelemetry-exporter-otlp-proto-http

Setup

from agno.agent import Agent
from opentelemetry import trace as trace_api
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor
from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
from openinference.instrumentation.agno import AgnoInstrumentor

provider = TracerProvider()
provider.add_span_processor(BatchSpanProcessor(OTLPSpanExporter(
    endpoint="https://api.zespan.com/v1/traces",
    headers={"x-api-key": "zsp_..."},
)))
trace_api.set_tracer_provider(provider)

AgnoInstrumentor().instrument(tracer_provider=provider)

# Every agent run below is now traced automatically
agent = Agent(instructions="You are a helpful assistant.")
response = agent.run("What's 2+2?")

What gets captured

Individual agent runs and multi-agent team coordination both map through the standard OpenTelemetry GenAI semantic conventions Zespan reads — a team’s parallel or sequential agent calls show up as one connected trace rather than disconnected runs.

Agno also ships a native OpenLIT integration, which is also OTel-based under the hood and can be pointed at Zespan the same way — use whichever instrumentor you already have configured.