GitHub radar
A playbook for building a profitable app with an AI agent
This playbook turns one developer's experience scaling 5 apps past $100k/month into a step-by-step guide — from picking an idea to your first sales.
A text guide with 14 chapters: how to pick an app niche, build it on a specific tech stack, and get it to revenue. No code — just instructions and starter file templates.
When it helps
Useful if you already have an app idea and want a ready-made plan plus a second opinion on which technologies to use. Don't follow it literally if you've already picked a different stack — you'll end up redoing work.
Pros
- Breaks the path from idea to revenue into 14 clear phases
- Gives ready starter file templates (worker, auth, config)
- Shows cost estimates at different scales — 1,000 to 1M downloads
- Suggests a cheap default AI model for simple repetitive tasks
Cons
- Unconventional stack: Cloudflare Workers, Better Auth, PlanetScale instead of usual tools
- Explicitly tells you to avoid Vercel, Supabase, Lovable, Bolt — popular beginner choices
- Cost comparison ignores differences in AI model quality
How to set it up — step by step
- 1Clone the repo at https://github.com/blakeandersonw/ai-app-playbook into an empty folder.
- 2Open your agent (Claude Code or similar) in that folder and ask it to read AGENTS.md from the root.
- 3Send the agent the text from the block below so it walks through chapter 00 and helps pick your niche.
- 4Ask the agent to explain the next chapters (backend, auth, AI, payments) for your specific app.
- 5Compare the suggested stack to what you already use — weigh whether switching is worth the rework.
- 6Check the cost table in chapter 12 to estimate expenses as your user count grows.
Text for your agent
Copy this and send it to your agent — Claude Code, Codex, any of them:
Clone https://github.com/blakeandersonw/ai-app-playbook and read AGENTS.md at the root — then follow it as the playbook instructs: walk me through chapter 00 (finding a niche) for my own app idea, and tell me what to do next on the stack.
▌ More finds