LLM Service

Where submissions are sent to be assessed, and the confidence at which they are published.

Changes here are written to the audit log against your name.

Endpoint

Service URL An OpenAI-compatible endpoint, such as a Synergia instance. API key Checking… Timeout Seconds a single call may take before it is abandoned.

Models

Classification model Scores whether a submission belongs here, and assigns its subjects. Translation model Translates the historical texts.

Thresholds

Confidence runs from 0 to 100. Starting values, meant to be tuned against real traffic.

Publish above Published without a human reading it first. Send to a curator above Below this, a submission is dropped as spam.