The novelist's drafting loop
Auctorly reads your manuscript, returns a real editor's letter, and rewrites the chapters that need work — so your next draft starts where the last one left off. Iterate as many times as your novel needs.
Why not just use ChatGPT
A general-purpose chatbot will tell you your book is great, surface 50 small notes nobody reads, miss the structural problems entirely — and then leave the rewriting to you. Three things make Auctorly different.
01
Four critics run in parallel — structure, character, prose, continuity — each with its own role prompt and forced output schema. ChatGPT runs a single pass and tries to do everything; Auctorly runs four sharp passes and synthesizes them.
02
Each critic plays a specific editorial persona — a senior editor who has rejected the manuscript before — with an explicit cap on praise. Findings are returned as ranked, scene-referenced JSON. Generic chatbots default to encouragement; this one is tuned to refuse it.
03
A synthesis pass dedupes findings across the four critics, ranks by severity × frequency, and produces a 5–10 issue prioritized letter. What a real editorial assessment looks like — not 200 pages of feedback no one reads.
How it works
Auctorly parses chapter breaks, indexes scenes, and stores the book privately under your account. Only you can see it.
No retrieval shortcuts. Each critic sees the whole book in context — that's how cross-chapter problems get caught. Roughly five to ten minutes end-to-end.
Five to ten ranked issues, each with diagnosis, scene references, and a suggested direction. Drill into the raw findings per critic when you want detail.
Generate a revised draft from the letter and run the whole pipeline again. Drafts are first-class — each one gets its own letter, and you can diff old against new chapter by chapter.
Sign in with Google. Upload your manuscript. Get critiqued, get rewritten, and decide where to go from there.