Providers

Auto-instrumentation for LLM providers.

Risicare automatically instruments popular LLM providers with zero code changes.

Supported Providers

Python SDK

JavaScript SDK

The JS SDK supports the same 12 native providers via explicit patchX() calls, plus 8 host-detected providers via patchOpenAI():

Additional JS providers (Google, Mistral, Groq, Cohere, Together, Ollama, HuggingFace, Cerebras, Bedrock) follow the same pattern — import from risicare/<provider> and call patchX(client). See the JS SDK reference for the full list.

How It Works

When you call risicare.init(), the SDK automatically patches supported provider libraries:

import risicare
from openai import OpenAI
 
risicare.init()
 
client = OpenAI()
 
# This call is automatically traced
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello!"}]
)

Captured Data

For each LLM call, Risicare captures:

Field	Description
`model`	Model name (gpt-4o, claude-3-sonnet, etc.)
`provider`	Provider name (openai, anthropic, etc.)
`prompt_tokens`	Input token count
`completion_tokens`	Output token count
`total_tokens`	Total token count
`cost_usd`	Estimated cost in USD
`latency_ms`	Request duration
`temperature`	Sampling temperature
`max_tokens`	Maximum output tokens
`prompt`	Input prompt (if `trace_content=True`)
`completion`	Output completion (if `trace_content=True`)

Manual Instrumentation

If auto-instrumentation doesn't meet your needs:

from risicare import trace
 
@trace(name="custom-llm-call", kind="llm_call")
def call_custom_llm(prompt: str) -> str:
    # Your custom LLM call
    return response

You can also set provider and model attributes manually:

from risicare import get_tracer
 
tracer = get_tracer()
 
with tracer.start_span("custom-llm", kind="llm_call") as span:
    span.set_attribute("llm.provider", "custom")
    span.set_attribute("llm.model", "my-model")
    response = call_llm(prompt)