From Terminal to Desktop: How a Desktop App Uses Copilot CLI as Its Entire AI Layer

This is a submission for the GitHub Copilot CLI Challenge

What I Built

I read Hacker News a lot. Tabs pile up, threads go deep, and it can be a bit of work to scroll through the stream of posts to find the ones that really match my interests. So I built something to fix this.

HN Reader is a WinUI 3 desktop app that uses GitHub Copilot CLI as its AI backbone to deliver two features I genuinely wanted:

• News Digest — The app fetches 50 popular HN stories from the last 24 hours and saves this data in the knowledge vault. A Copilot CLI agent then reads this data, calls tools to enrich stories with additional context, and generates a personalized digest grouped by your configured interests (e.g., "AI/ML", "Startups", "Systems Programming"). Each group includes ranked stories with concise summaries so you can scan a full day of HN in just a few minutes.

• Story Insights — Select any story and get a structured breakdown: TL;DR, core argument, discussion map, points of agreement/disagreement, and practical takeaways.

By building on the GitHub Copilot SDK, the app does not require any LLM API keys. Anyone with Copilot CLI installed gets the AI features out of the box: no BYOK, no env vars.

---

The News Digest view — stories grouped by user interests with AI-generated summaries

---

An expanded interest group showing ranked stories and their summaries

---

Story Insight panel

---

Knowledge Vault configuration — local folder where the app stores data for the agent to work with

---

User interests configuration

---

App Architecture

---

Demo

GitHub Repo: HN Reader

My Experience with GitHub Copilot CLI

This project is built around one central idea: a terminal-native AI assistant can transform a desktop experience. Copilot CLI isn't just a developer tool; it can also serve as the AI layer for an app. Here, Copilot SDK acts as the perfect interface layer, providing a clean bridge between the app and Copilot CLI capabilities.

Using GitHub Copilot CLI made me faster, but more importantly, it made the app better: more helpful, more personalized, and more valuable for real users.