Getting Started

Getting Started

Install (End Users)

Download the latest installer from GitHub Releases:

• .msi — standard Windows Installer with Start Menu shortcuts and Add/Remove Programs entry (~28 MB)
• -setup.exe — portable NSIS installer, runs directly without Start Menu entries (~27 MB)

Both bundle the Python backend as a sidecar. No Python installation required on the target machine.

Open-Write is not yet code-signed. Windows SmartScreen may show a warning. Click More info, then Run anyway.

System Requirements

• Windows 10 or 11
• ~60 MB free disk space for the installer
• An AI provider API key (optional — local models via LM Studio or Ollama work without one)

First Run

1. Launch Open-Write
2. Open Settings (gear icon)
3. Pick a model from the Recommended Models panel — this works without any API key configured, or paste your own provider API key
4. Click New Project on the home screen and choose a folder

Your project is just a folder of Markdown files. You can back it up, sync it to a cloud drive, or commit it to a personal git repo.

Project Structure

Each project is a folder you own:

MyNovel/
  project.json                     project settings
  manuscript/                      chapter .md files (one per file)
    01-chapter-one.md
    02-chapter-two.md
  notes/                           outline, style guide, themes
    outline.md
    style-guide.md
  profiles/                        character, relationship, location, lore profiles
    characters/
    relationships/
    locations/
    lore/
  summaries/                       chapter and scene summaries
    chapters/
    scenes/
  exports/                         combined manuscript + dated snapshots
  .open-write/                     local cache (safe to delete, rebuilt from Markdown)

For pipeline projects, additional folders are created:

  bible/                           concept, outline, format rules
  state/                           pipeline run state, completion manifest
  manuscript/chapters/             per-chapter drafts
  critic_outputs/                  critic review artifacts
  coverage_reports/                editorial coverage

Configuring AI Providers

Open-Write supports 22 LLM providers. To configure one:

1. Open Settings (gear icon)
2. Scroll to LLM Providers
3. Find your provider and paste your API key
4. Click Test to verify the connection
5. Select a model from the Recommended Models panel or type one manually

For local inference (no API key needed):

• LM Studio — default endpoint http://localhost:1234/v1
• Ollama — default endpoint http://localhost:11434/v1

See LLM Providers for the full list.

Updates

The auto-updater is not yet configured. Check GitHub Releases for new versions.