# OpenRouter launches the Extractions API in alpha for pulling typed fields from documents

> Original title: Introducing the Extractions API

- Source: OpenRouter Blog
- Published: 2026-11-10T00:00:00.000Z
- Bad Influence: https://badinfluence.news/items/j3dny7r8iz9rhq212hjszc5ov
- Original: https://openrouter.ai/blog/announcements/extractions-api/

## Summary

OpenRouter launches the Extractions API in alpha, which pulls named, typed fields such as totals, dates, and contacts from text, images, and PDFs. Clairic returns spans, p_null, and confidence for single values, while Extract returns a confidence per value and no p_null. The launch includes three models from Cloudflare, Inception, and ScaleDown, and the source says output is free on all three.

## Why it matters

The post shows how a field-typed request returns values with spans, p_null, and confidence, so readers can check where each answer came from.