hallutraceai 0.1.4

Python SDK for HalluTrace AI - LLM hallucination detection & tracing

HalluTrace AI - Python SDK

Detect hallucinations in your LLM applications with HalluTrace AI.

Installation

pip install hallutraceai

Quick Start

from hallutraceai import HalluTrace

ht = HalluTrace(api_key="sk_live_your_api_key")

# Send a trace
ht.trace(
    session_id="chat-123",
    message_id="msg-1",
    type="agent",
    input="What is the capital of France?",
    output="The capital of France is Paris.",
    system_prompt="You are a helpful assistant.",
    model_name="gpt-4",
)

Batch Tracing

ht.trace_batch([
    {
        "session_id": "chat-123",
        "messages": [
            {
                "message_id": "msg-1",
                "type": "agent",
                "input": "Hello",
                "output": "Hi! How can I help?",
                "model_name": "gpt-4",
            },
            {
                "message_id": "msg-2",
                "type": "tool",
                "input": "search: weather",
                "output": '{"temp": 72, "condition": "sunny"}',
            },
        ],
    }
])

Message Types

• agent — LLM responses (evaluated for hallucinations)
• tool — Tool/function call results (logged but not evaluated)

API Reference

HalluTrace(api_key, base_url=None, timeout=30.0)

Create a client instance.

ht.trace(session_id, message_id=None, type="agent", input="", output="", ...)

Send a single trace. Returns session and message info.

ht.trace_batch(traces)

Send multiple traces in one request.

Links

• Dashboard
• API Docs
• Support