Trace PydanticAI with BotGauge

Records PydanticAI agent runs. PydanticAI instruments itself, so the in-code sample turns that on and adds OpenInference’s span processor.


Extra	`[pydantic-ai]`
Verified framework versions	pydantic-ai 1.25.1 and 2.57.0
Instrumentor installed by the extra	`openinference-instrumentation-pydantic-ai 0.1.30`
Zero-code run	Traced

Also verified: pydantic-ai 1.25.1 with instrumentor 0.1.27. The extra no longer installs that older instrumentor. See supported frameworks for how verification works.

In code

Call tracing.init(), attach the framework instrumentor to the provider it returns, and wrap the work in tracing.turn() and tracing.agent().

The install line uses PydanticAI 2.57, which needs Python 3.11 or newer and no other pins.

Python 3.10: use PydanticAI 1.25

On Python 3.10, use the tested PydanticAI 1.25 instead. It needs OpenTelemetry 1.39 and genai-prices older than 0.1 to record token counts, because PydanticAI before 2.17 records none with genai-prices 0.1.0 or newer, and the trace would then show zero cost. In code, pass version=3 to InstrumentationSettings on PydanticAI 1.x and version=5 on 2.x.

terminal

pip install "botgauge[pydantic-ai]" pydantic-ai==1.25.1 \
  opentelemetry-api==1.39.1 opentelemetry-sdk==1.39.1 \
  opentelemetry-exporter-otlp-proto-http==1.39.1 genai-prices==0.0.73

Install

terminal

pip install "botgauge[pydantic-ai]" pydantic-ai==2.57.0

Add the code

app.py

import os

from botgauge import tracing
from openinference.instrumentation.pydantic_ai import OpenInferenceSpanProcessor
from pydantic_ai import Agent
from pydantic_ai.models.instrumented import InstrumentationSettings
from pydantic_ai.models.openai import OpenAIChatModel
from pydantic_ai.providers.openai import OpenAIProvider

provider = tracing.init()
Agent.instrument_all(InstrumentationSettings(tracer_provider=provider, version=5))
provider.add_span_processor(OpenInferenceSpanProcessor())

openai_provider = OpenAIProvider(
    base_url=os.environ["OPENAI_BASE_URL"], api_key=os.environ["OPENAI_API_KEY"]
)
model = OpenAIChatModel(os.environ["OPENAI_MODEL"], provider=openai_provider)
agent = Agent(model, name="assistant")

with tracing.turn(session_id="ses_pydantic_ai"), tracing.agent("pydantic-ai-agent"):
    result = agent.run_sync("In one sentence, what is OpenTelemetry?")
print(result.output)

tracing.shutdown()

Run

terminal

python app.py

Needs BOTGAUGE_API_KEY, OPENAI_BASE_URL, OPENAI_API_KEY, OPENAI_MODEL.

Zero-code

Leave your app unchanged and start it with botgauge instrument. The sample still wraps the work in tracing.turn() and tracing.agent() so spans carry session and agent ids.

The install line uses PydanticAI 2.57, which needs Python 3.11 or newer and no other pins.

Python 3.10: use PydanticAI 1.25

On Python 3.10, use the tested PydanticAI 1.25 instead. It needs OpenTelemetry 1.39 and genai-prices older than 0.1 to record token counts, because PydanticAI before 2.17 records none with genai-prices 0.1.0 or newer, and the trace would then show zero cost. botgauge instrument picks the right instrumentation version for you.

terminal

pip install "botgauge[pydantic-ai]" pydantic-ai==1.25.1 \
  opentelemetry-api==1.39.1 opentelemetry-sdk==1.39.1 \
  opentelemetry-exporter-otlp-proto-http==1.39.1 genai-prices==0.0.73

Install

terminal

pip install "botgauge[pydantic-ai]" pydantic-ai==2.57.0

Add the code

app.py

import os

from botgauge import tracing
from pydantic_ai import Agent
from pydantic_ai.models.openai import OpenAIChatModel
from pydantic_ai.providers.openai import OpenAIProvider

openai_provider = OpenAIProvider(
    base_url=os.environ["OPENAI_BASE_URL"], api_key=os.environ["OPENAI_API_KEY"]
)
model = OpenAIChatModel(os.environ["OPENAI_MODEL"], provider=openai_provider)
agent = Agent(model, name="assistant")

with tracing.turn(session_id="ses_pydantic_ai"), tracing.agent("pydantic-ai-agent"):
    result = agent.run_sync("In one sentence, what is OpenTelemetry?")
print(result.output)

Run

terminal

botgauge instrument python app.py

Needs BOTGAUGE_API_KEY, OPENAI_BASE_URL, OPENAI_API_KEY, OPENAI_MODEL.

View PydanticAI spans live

Watch your PydanticAI agent runs with token counts appear in BotGauge.

Try BotGauge →

See versions not verified if you run a different version, capture and redaction to control exported content, and troubleshooting if spans are missing.

Frameworks

More

Trace PydanticAI with BotGauge

In code

Zero-code

​In code

​Zero-code

​Related

In code

Zero-code

Related