Skip to main content
LLM provider wrappers are the simplest way to add tracing to your LLM calls. Wrap your provider client once and every API call is automatically traced — no other code changes required. The wrapper returns the same client type with tracing injected. Your existing code continues to work identically. All wrappers automatically capture: input messages, output messages, model name, token usage, model parameters, streaming support, time-to-first-token for streaming, and error details.

​
Comparison

WrapperInstall ExtraProvider Methods Traced
wrap_openai"pandaprobe[openai]"chat.completions.create, responses.create
wrap_anthropic"pandaprobe[anthropic]"messages.create, messages.stream
wrap_gemini"pandaprobe[gemini]"models.generate_content, models.generate_content_stream
wrap_mistral"pandaprobe[mistral]"chat.complete, chat.complete_async, chat.stream, chat.stream_async
wrap_bedrock (beta)"pandaprobe[bedrock]"converse, converse_stream, invoke_model, invoke_model_with_response_stream
Install only the provider extras you use so dependency resolution stays minimal.

​
Quick example

from pandaprobe.wrappers import wrap_openai
from openai import OpenAI

client = wrap_openai(OpenAI())
# Use client exactly as before — all calls are traced automatically
Wrappers work seamlessly with manual instrumentation. If a wrapper call happens inside a pandaprobe.start_trace() or @pandaprobe.trace context, the LLM span is automatically nested as a child span.

​
Provider guides

OpenAI

Chat Completions and Responses API, streaming, and tool spans

Anthropic

Messages API, streaming patterns, and extended thinking

Google Gemini

generate_content, async, streaming, and thinking mode

Mistral AI

Chat completions, async, and streaming

AWS Bedrock (beta)

Converse and InvokeModel APIs, streaming, and aioboto3 support