Fast Chapter

Voice-first desktop app for drafting books chapter by chapter.

Download

• Latest desktop builds: GitHub Releases
• CI artifacts (before release): Build workflow runs

What Fast Chapter Is

Fast Chapter is an Electron app for writers who want to draft by voice first and refine later. It is local-first by default and designed around practical book production workflows.

Core capabilities

• Local user profiles and local project storage
• Bookshelf + per-book workspace
• Automatic LaTeX scaffold for new books
• Explorer actions: create, rename, move, delete, upload
• Multi-select explorer actions (bulk move/delete)
• Recording save flow + organized folder structure
• Background transcription jobs with gpt-4o-transcribe
• Write-book workflow with Codex session support
• LaTeX compile + in-app PDF preview (main.tex)
• Export books as ZIP archives

Quick Start (Developers)

Prerequisites

• Node.js 20+
• npm 10+
• Optional for PDF compile preview: latexmk (recommended) or pdflatex on PATH

Install

npm install

Run in development

npm run dev

Useful split commands:

npm run dev:ui
npm run dev:electron

Build and run production locally

npm run build
npm run start

Desktop Packaging

npm run dist
npm run dist:win
npm run dist:mac

Artifacts are written to release/.

• Windows installer: *-setup.exe
• Windows portable: *-portable.exe
• macOS installer: *.dmg
• macOS archive: *.zip

Project Structure

• Electron main/preload: electron/main.cjs, electron/preload.cjs
• Renderer: src/* (React 18 + Vite + TypeScript)
• UI primitives: src/components/ui/*
• Prompt templates: prompts/*

Book Scaffold Rules

When a new book is created, Fast Chapter ensures:

• book.json
• main.tex
• cover-page.tex
• back-page.tex
• chapters/chapter-1/chapter-1.tex
• chapters/chapter-1/assets/
• recordings/
• transcriptions/

main.tex is the project entry file and includes cover page, table of contents, chapter includes, and back page.

Recording/transcription files are organized under:

• recordings/initial-outline/
• recordings/chapters/chapter-N/
• recordings/miscellaneous/
• transcriptions/initial-outline/
• transcriptions/chapters/chapter-N/
• transcriptions/miscellaneous/

OpenAI Transcription

• API key is stored in local user profile (never hardcoded)
• Endpoint: POST /v1/audio/transcriptions
• Model: gpt-4o-transcribe
• UI tracks job state: queued, in_progress, completed, failed
• Upload limit: 25 MB per file
• Supported formats: mp3, mp4, mpeg, mpga, m4a, wav, webm

CI and Releases

• Workflow: .github/workflows/build-desktop.yml
• Builds Windows and macOS artifacts
• Uploads build artifacts to workflow runs
• Publishes release assets on v* tags or manual dispatch

Contributing

• Guide: CONTRIBUTING.md
• Security policy: SECURITY.md
• Code of conduct: CODE_OF_CONDUCT.md
• Contributors: CONTRIBUTORS.md

License

Licensed under GNU Affero General Public License v3.0 (AGPL-3.0).

• Full text: LICENSE