An AI-powered writing assistant for macOS. Spellcheck, rewrite, and continue your text using any LLM via OpenRouter.
macOS — requires macOS 12+
| Architecture | Link |
|---|---|
| Apple Silicon (M1/M2/M3) | Pero Editor-0.1.2-arm64.dmg |
| Intel (x64) | Pero Editor-0.1.2.dmg |
The app includes auto-update — new versions install automatically.
- Rich text editor powered by Tiptap / ProseMirror
- AI suggestions: spellcheck, rewrite in 4 styles, continue writing
- Real-time sync with version conflict resolution (WebSocket)
- Author post library — sidebar with all your posts
- Desktop & mobile preview modes
- Bring your own API key — any model supported by OpenRouter
The project is a monorepo with three apps:
app-frontend/ React 19 + Vite + Tiptap + Tailwind CSS
app-backend/ Node.js + Moleculer + WebSocket + Prisma + SQLite
app-desktop/ Electron — packages frontend + backend into a macOS app
The desktop app bundles both the frontend and backend — no external server needed.
- Node.js 20+
- npm
cd app-backend
cp .env.example .env
# Generate a secure encryption key:
# openssl rand -hex 32
# Paste the result as AI_SECRETS_ENCRYPTION_KEY in .env
npm install
npm run db:migrate
npm run devcd app-frontend
cp .env.example .env.local
# Set VITE_WS_URL=ws://localhost:8080
npm install
npm run devOpen http://localhost:3000 in your browser.
Requires the backend to be built first:
cd app-backend
npm run build
cd ../app-desktop
npm install
npm run devTo build the DMG:
cd app-desktop
npm run distThe app connects to OpenRouter to access LLMs. Configure your API key and preferred model in the AI Settings modal inside the app (gear icon).
Supported: any model available on OpenRouter (GPT-4o, Claude, Gemini, Llama, etc.)
app-backend/.env
AI_SECRETS_ENCRYPTION_KEY= # 32-byte hex key: openssl rand -hex 32app-frontend/.env.local
VITE_WS_URL=ws://localhost:8080
VITE_WORKSPACE_ID=default-workspace
VITE_AUTHOR_USER_ID=default-user
VITE_POST_LIST_LIMIT=20See CONTRIBUTING.md.