Skip to main content

​
Prerequisites

Install the Tracia SDK: 
pip install tracia
Set your API keys as environment variables:
.env
TRACIA_API_KEY=tr_your_tracia_key
OPENAI_API_KEY=sk-your-openai-key
ANTHROPIC_API_KEY=sk-ant-your-anthropic-key
GOOGLE_API_KEY=your_google_key
The Python SDK uses LiteLLM under the hood. LiteLLM is included as a dependency and handles all provider communication.

​
OpenAI

from tracia import Tracia

client = Tracia(api_key="tr_your_api_key")

result = client.run_local(
    model="gpt-4o",
    messages=[
        {"role": "system", "content": "You are a helpful coding assistant."},
        {"role": "user", "content": "Write a function to reverse a string in Python."},
    ],
    temperature=0.7,
    max_output_tokens=500,
)

print(result.text)
print(f"Tokens used: {result.usage.total_tokens}")

​
Anthropic

from tracia import Tracia

client = Tracia(api_key="tr_your_api_key")

result = client.run_local(
    model="claude-sonnet-4-20250514",
    messages=[
        {"role": "system", "content": "You are a creative writing assistant."},
        {"role": "user", "content": "Write a short story opening about a time traveler."},
    ],
    temperature=0.9,
    max_output_tokens=1000,
)

print(result.text)
print(f"Provider: {result.provider}")

​
Google

from tracia import Tracia

client = Tracia(api_key="tr_your_api_key")

result = client.run_local(
    model="gemini-2.0-flash",
    messages=[
        {"role": "user", "content": "Explain the difference between HTTP and HTTPS."},
    ],
    temperature=0.5,
)

print(result.text)
print(f"Latency: {result.latency_ms}ms")

​
Multi-Turn Conversations

Include previous messages to maintain conversation context: 
result = client.run_local(
    model="gpt-4o",
    messages=[
        {"role": "system", "content": "You are a math tutor."},
        {"role": "user", "content": "What is 15% of 80?"},
        {"role": "assistant", "content": "15% of 80 is 12."},
        {"role": "user", "content": "How did you calculate that?"},
    ],
)

​
With Tracing Metadata

Add tags and user identifiers for filtering in the Tracia dashboard: 
result = client.run_local(
    model="gpt-4o",
    messages=[
        {"role": "user", "content": "Summarize the key points of agile development."},
    ],
    tags=["production", "summarization"],
    user_id="user_abc123",
    session_id="session_xyz789",
)

print(f"Span ID: {result.span_id}")
# View this span in the Tracia dashboard

​
Streaming

Enable streaming to receive responses in real-time: 
stream = client.run_local(
    model="gpt-4o",
    messages=[
        {"role": "user", "content": "Write a short poem about coding."},
    ],
    stream=True,
)

# Span ID is available immediately
print(f"Span: {stream.span_id}")

# Iterate to receive chunks
for chunk in stream:
    print(chunk, end="")

# Get final result with usage stats
result = stream.result.result()  # Future[StreamResult] → StreamResult
print(f"\nTokens: {result.usage.total_tokens}")
See Streaming for more details.

​
Async Usage

Use arun_local() for async contexts: 
result = await client.arun_local(
    model="gpt-4o",
    messages=[
        {"role": "user", "content": "Hello!"},
    ],
)
print(result.text)

​
Without Tracing

Disable tracing when you don’t need observability: 
result = client.run_local(
    model="gpt-4o-mini",
    messages=[
        {"role": "user", "content": "What is 2 + 2?"},
    ],
    send_trace=False,
)

print(result.span_id)  # "sp_..." (still populated, just not sent to Tracia)