The Dispatch

The Dispatch is a local-first, AI-native desktop news reader. It runs on your computer, pulls stories from your sources, notify you when there's a breaking news, and generates daily digests.

Capabilities

• Agents that fetch articles from any website sources.
• LLM generates summary, tags, grading, key points for each article.
• Model/provider management (supports Anthropic and OpenAI-compatible providers). Per-task model routing.
• Scheduled fetch, articles processing, and digest jobs with cron config.
• IM notifications (supports Telegram).

Installation

Download the latest release for your platform from Releases:

Platform	Format
macOS	.dmg
Windows	.exe installer
Linux	.AppImage

Getting Started

1. Open The Dispatch after installing.
2. The onboarding wizard will guide you through setting up an LLM provider (Anthropic, OpenAI, or any OpenAI-compatible endpoint such as Ollama or LM Studio).
3. Select which models to use for each task (summarization, classification, grading, etc.).
4. Add your first news source — an RSS feed URL or any web page.

The Dispatch will begin fetching and processing articles.

Monorepo layout

• apps/desktop: Electron + React + Vite UI
• packages/server: Hono + tRPC server and background jobs
• packages/db: SQLite + Drizzle schema/migrations
• packages/lib: shared LLM/provider logic
• packages/api: shared API types

Development

1. Install dependencies:

pnpm install

2. Create your local database schema:

pnpm --filter @dispatch/db migrate

3. (Optional) Copy env template:

cp sample.env .env

4. Start:

pnpm dev

pnpm dev runs the desktop app and local server via Turborepo.