Latence TRACE Python SDK

Real-time safety for knowledge agents. Groundedness verification, privacy redaction, and context compression — without replacing your stack.

PyPI · Quickstart · Integrations · Examples · Website

pip install latence

from latence import Latence

trace = Latence(api_key="lat_...")

score = trace.grounding.rag(
    query="Can I promise this customer a refund?",
    response_text="Yes, the refund will arrive within 48 hours.",
    raw_context="Refunds require manual finance approval before timelines are promised.",
)

print(score.risk_band)
print(score.runtime_decision)

What TRACE Does

• Groundedness — verify RAG answers against retrieved context.
• Compression — reduce long context to what matters.
• Privacy — redact private data before it spreads through tools, logs, or prompts.

Getting Started

export LATENCE_TRACE_API_KEY="lat_..."

from latence import Latence

trace = Latence()

The SDK connects to https://api.latence.ai by default. Override with base_url or LATENCE_TRACE_URL.

API Surface

# Groundedness scoring
score = trace.grounding.rag(query=..., response_text=..., raw_context=...)

# Privacy redaction
redacted = trace.privacy.redact(text=...)

# Context compression
compressed = trace.compression.text(text=..., compression_rate=0.4)

Latence is synchronous. AsyncLatence exposes the same surface for asyncio services.

Base dependencies are only httpx and pydantic.

Integrations

Direct calls are the recommended path. Optional helpers live under latence.integrations.

pip install "latence[langchain]"
pip install "latence[llama_index]"
pip install "latence[openai]"
pip install "latence[haystack]"

Docs: Integrations

More

• Quickstart
• Integrations
• Errors
• Publishing
• Examples
• Website