Cursor-style AI document editor with diff view
Problem
People using AI for writing documents have no way to see what the AI changed: tools rewrite whole paragraphs with no diff view, no way to accept/reject individual edits, and formatting/styling still has to be fixed manually. Generating PDFs/DOCX via Claude or Codex is clunky compared to typing in a real editor.
Opportunity
A word processor with an AI copilot layer that proposes edits inline as reviewable diffs (accept/reject per change) plus auto-formatting and styling commands — essentially Cursor's editing experience applied to documents.
Market analysis
The pain is real and widely shared, but the core UX already exists: Athens markets itself as 'Cursor for writing' with inline AI diffs and per-change accept/reject, while Zoho Writer and Google Docs bundle AI suggestions into familiar track-changes flows. What remains unclaimed is the boring half of the request — automatic DOCX/PDF formatting and styling — which is also the hardest engineering lift.
Market · Knowledge workers, consultants, and students who draft with AI assistance; strong recurring demand signal on HN, but mainstream users are already absorbed by Google Workspace's bundled AI.
Pricing · Freemium editors are the norm: Zoho Writer is free, Gemini-in-Docs arrives with Workspace plans from roughly $8/user/mo, Google AI Pro at $19.99/mo — a solo product must live in the $10-20/mo Pro tier.
Pros
- + Genuine, recurring pain for anyone writing with AI assistance.
- + Diff-based review is a proven pattern (Cursor, track changes); the UX model is validated.
- + Auto-formatting to clean DOCX/PDF output is a concrete, demo-able wedge nobody owns end-to-end.
Cons
- − Athens already ships the exact pitch: markdown WYSIWYG with inline AI diffs and per-change accept/reject.
- − Google and Zoho can bundle equivalent AI suggestions into editors users already live in.
- − Full DOCX/PDF fidelity plus a real editor is a multi-year engineering lift for a solo builder.
Existing / similar tools
Source
Hacker News (Ask HN)
The editor itself is the wrong battle: Athens proves the Cursor-for-writing UX can be built, and incumbents will keep absorbing the chat-to-dictation flow. The defensible wedge is the layer nobody wants to build — deterministic document formatting (styles, numbering, TOC, DOCX/PDF export that survives real-world templates). Sell that as an engine that plugs into existing editors or agent pipelines first; a full word processor only makes sense once the formatting engine is the moat.