New Project Ideas — What to Build to Fill the Gaps
⏳ The Problem
Your GitHub has strong projects, but there are narrative gaps. The portfolio tells the story of a full-stack + systems developer, but some categories are weak or missing. These project ideas fill those gaps while being genuinely interesting to build.
Priority: Build ONE of these well, rather than three of them poorly.
🔮 Idea 1: dotbrowser — A Terminal-Rendered Dotfile Viewer ★★★★★
Category gap filled: Developer tooling + Open source potential
What it is: A web app where you paste a GitHub username and it renders their dotfiles beautifully — syntax highlighted, grouped by category (shell, editor, WM), with a visual “setup score” or “ricing grade.”
Why build it:
- Directly related to your interests (dotfiles, Linux, ricing)
- Could get stars — the Linux community loves this stuff
- Demonstrates GitHub API, syntax highlighting, and design skills
- Perfect project to showcase ON the portfolio (meta-recursive)
Tech: Next.js, TypeScript, shadcn/ui, Shiki, GitHub API
Effort: 2-3 weeks
🔮 Idea 2: typecheck — A TypeScript Playground with Constraint Puzzles ★★★★☆
Category gap filled: TypeScript expertise showcase + Educational content
What it is: An interactive playground where developers solve TypeScript type puzzles. Given a set of test cases and a type signature, write the correct type. Like type-challenges but with a beautiful UI, hints, and a progression system.
Why build it:
- You say you love TypeScript. This PROVES it at a deep level.
- Educational tools get shared widely (good for stars/visibility)
- Demonstrates advanced TS knowledge (generics, conditional types, mapped types)
- Interactive web app shows frontend skill
Tech: Next.js, TypeScript, Monaco Editor, shadcn/ui
Effort: 3-4 weeks
🔮 Idea 3: patchwork — A TUI Git Diff Viewer in Rust ★★★★☆
Category gap filled: Rust CLI tool that people might actually use
What it is:
A terminal application (TUI, using ratatui) that shows git diffs
beautifully — side-by-side, syntax highlighted, with keyboard navigation.
Like delta but more visual, maybe with inline comments.
Why build it:
- Rust TUI app — extends your 4at and seroost narrative
- Developer tools get stars if they solve a real pain point
- Demonstrates Rust, TUI development, and git internals
- You already know Rust TUI development from mdt
Tech: Rust, ratatui, tree-sitter (for syntax highlighting), git2-rs
Effort: 4-6 weeks
🔮 Idea 4: bento — A Personal Dashboard / New Tab Page ★★★☆☆
Category gap filled: Design-forward web app + Daily utility
What it is: A customizable browser new-tab page with widgets: GitHub activity, weather, bookmarks, quick links, Spotify now-playing, system status. Tokyo Night themed. Exports/imports config as JSON.
Why build it:
- Visual, design-heavy project that showcases UI skills
- Daily-use tool (you’d actually use it)
- Widget architecture demonstrates component composition
- Can integrate with APIs you already know (Spotify, GitHub)
Tech: Next.js or plain React, TypeScript, shadcn/ui, browser extension API
Effort: 2-3 weeks
🔮 Idea 5: shellshare — Share Terminal Sessions via Web ★★★★★
Category gap filled: Full-stack + Real-time + Systems (the trifecta)
What it is: Record your terminal session (like asciinema) and share it via a URL. Others can watch the replay in the browser with full formatting. Maybe even live-stream terminal sessions.
Why build it:
- Combines your web + systems duality perfectly
- Terminal recording is niche but useful (conference talks, tutorials)
- Real-time streaming shows WebSocket/infrastructure skills
- Could actually get traction as an indie tool
Tech:
- Backend: Rust (recording daemon) or Go
- Frontend: Next.js, xterm.js, WebSocket
- Storage: S3/R2 for recordings
Effort: 4-6 weeks
🔮 Idea 6: keybind — Interactive Keyboard Shortcut Trainer ★★★☆☆
Category gap filled: Fun/creative web project
What it is: An interactive web app that teaches keyboard shortcuts for different tools (Vim, VS Code, Hyprland, tmux). Users see a prompt, press the correct keys, and get scored. Leaderboard optional.
Why build it:
- Fun, shareable, and viral potential
- Related to your CLI/keyboard-driven workflow aesthetic
- Quick to build and polish
- Good for blog content (“I built a keyboard shortcut trainer”)
Tech: Next.js, TypeScript, shadcn/ui, Motion
Effort: 1-2 weeks
✅ My Recommendation
Build ONE of these before launching the portfolio:
If you want stars and visibility → Idea 1 (dotbrowser)
It’s the most shareable, most related to your identity, and most likely to resonate with the Linux/ricing community.
If you want to flex TypeScript depth → Idea 2 (typecheck)
Proves you understand TypeScript at a level beyond “I use it for React.”
If you want the most impressive single project → Idea 5 (shellshare)
Full-stack + systems + real-time. This would be a Tier 1 hero project immediately. But it’s also the most work.
If you want something fast → Idea 6 (keybind)
Build it in a weekend. Ship it. Write a blog post. Move on.
✅ What NOT to Build
| Avoid | Why |
|---|---|
| Another todo app | Please. |
| Another chat app | You already have 4at. |
| Another dashboard | You already have ad-my-brand-insights. |
| A social media clone | Twitter/Instagram clones say nothing new. |
| An AI wrapper | ”I called the OpenAI API” is not a project. |
| A portfolio template | The portfolio itself is the project. |
⏳ Before Building Anything New
Fix your existing projects first. The ROI on adding a README, deploying to Vercel, and recording a demo GIF is much higher than starting a new project. Specifically:
- Deploy ad-my-brand-insights to Vercel (30 min)
- Record a GIF of mdt’s TUI in action (20 min)
- Record a demo video of musializer (10 min)
- Add screenshots to snippetbox README (15 min)
These four tasks will improve your portfolio more than a new project.