Mistral Introduces a Moderation API to Detect Undesirable Text Content

IBL News | New York

Paris-based AI startup Mistral launched a new API for content moderation last week. This addresses growing concerns about data privacy, latency, and compliance, which can prove especially attractive to European companies subject to strict data protection regulations.

This API powers the moderation service in Mistral’s Le Chat. Powered by a fine-tuned model (Ministral 8B), it can be tailored to specific applications and safety standards.

Mistral is releasing two end-points: one for raw text and one for conversational content.

The API is the same API that powers moderation in Mistral’s Le Chat chatbot platform.

It includes English, French, and German languages into one of nine categories: sexual, hate and discrimination, violence and threats, dangerous and criminal content, self-harm, health, financial, law, and personally identifiable information.

Mistral says the moderation API can be applied to raw or conversational text.

The moderation API is available through Mistral’s cloud platform, with pricing based on usage.

The release follows Mistral’s recent announcement of high-profile partnerships, including deals with Microsoft AzureQualcomm, and SAP.