Research Publisher

Turn evidence into living documents, PDFs, books, lessons, and narrated studies.

Scribe's publishing engine should generate from structured research, not loose AI prose. Every section, claim, confidence label, and citation can become part of the knowledge base.

1 Research Brief

User chooses topic, scope, audience, tone, language, and output formats.

2 Source Packet

Scribe gathers primary sources, books, papers, images, maps, and licensing notes.

3 Claim Ledger

Every factual claim gets receipts, confidence, competing views, and review status.

4 Publication Build

The same structured file renders as web article, PDF, EPUB, slides, and audio.

Prototype Publication Loading...

Topic Output Depth

Source Ingestion Source URL Source Type

Knowledge Library

Lesson Builder Lessons Minutes Each Audience Tone

Publish Result Public Title Subtitle Slug Public Summary Publication Type Visibility Evidence Status Tags Display Name Publish as Anonymous Image URL Image Alt Text Image License Image Source URL

Confirm public fields, claims, and image rights were reviewed

Publisher backend will create saved jobs and exports here.

This first backend creates jobs, validates claim receipts, and exports PDF. The agent research worker comes next.

Loading publication model...