Confidence Scoring Python Packages

confidence-escalation

Framework-agnostic confidence-gated escalation middleware for LLM agents: multi-signal scoring (logprob, verbalized, tool risk), threshold policies, and escalation handlers for LangChain, CrewAI, AutoGen, and Google ADK.

seapig

Uncertainty based selection of compatible inputs

confident-extract

Deterministic structured extraction from noisy LLM/OCR output. Zero LLM round-trips, microsecond latency, confidence score on every result. msgspec · Pydantic · dataclasses.

sacor

The Open Document Extraction Engine

murphy-confidence

MFGC confidence scoring and safety gates for AI agents. Zero dependencies.