Ideas • By John Iseghohi (opens in new tab) • May 1, 2026

7 Things You Can Build This Weekend with Claude Code

Each one is shippable in a single weekend. Each comes with a starter prompt you can paste straight in.

"What should I build?" is the wrong question. You don't need a big idea. You need a small one you can finish.

Below are seven projects sized to fit a Saturday and a Sunday. Each one is small enough that Claude Code can scaffold it in an hour and you can polish it in five. Pick one. Open your terminal. Ship it.

If you don't have Claude Code installed yet, start with our non-technical founder's guide first.

1

A personal CRM that lives in your browser

Why it matters: Most founders need to remember 50–200 people, not 5,000. HubSpot is overkill. A spreadsheet is too dumb.

What you ship: A single-page web app with a list of contacts, last-touched dates, follow-up reminders, and a notes field per person.

Starter prompt:
"Build a personal CRM. Single-page React app with Tailwind. Local storage only—no backend. Contacts have name, email, last contacted date, follow-up date, and a notes field. List view sorted by who's most overdue for follow-up. Add/edit modal. Export-to-CSV button. Dark theme."

Ship time: 4–6 hours.

2

A niche newsletter aggregator

Why it matters: Every niche has 5–15 great Substacks. Nobody wants to subscribe to all of them. A weekly digest is a real business model—Morning Brew, Milk Road, The Hustle.

What you ship: A page that pulls the latest post from a hand-picked list of RSS feeds, ranks them by recency, and shows the top 10 with summaries.

Starter prompt:
"Build a niche newsletter aggregator. Next.js app. Hardcoded list of 10 RSS feeds for [your niche]. Server-side fetch them on page load. Show title, source, date, and a 2-sentence AI summary using the Claude API. Email-signup form at the top stores to a Supabase table. Deploy to Vercel."

Ship time: 6–8 hours.

3

A Telegram or Discord bot for your community

Why it matters: If you run a community or a paid group, automating the boring parts (FAQs, onboarding, weekly threads) is one of the highest-leverage things you can do.

What you ship: A bot that answers FAQs from a knowledge base, posts a weekly intro thread, and pings new members.

Starter prompt:
"Build a Telegram bot in Node.js. It listens for /faq commands and answers from a markdown file in the repo. On a cron schedule, posts a weekly 'introduce yourself' thread. Greets new members with a custom message. Use the official Telegram Bot API. Deploy to Fly.io."

Ship time: 4–6 hours.

Want more ideas like these, with revenue research baked in?

Browse Startup Ideas
4

An internal tool you'd actually pay for at work

Why it matters: The most reliable wedge into B2B software is "I built this for my team and other teams want it." If it works for you, it works for someone else.

What you ship: Pick the worst spreadsheet at your job. Replace it with a small web app that does one thing better. PTO tracker. Vendor invoice approval. Weekly ops dashboard. Anything.

Starter prompt:
"Build a PTO request tracker. Next.js + Supabase auth. Employees log in, submit a request (start date, end date, reason). Managers see a dashboard with pending requests and can approve/deny. Email notifications via Resend. Mobile-responsive."

Ship time: 8–12 hours.

5

A landing page that validates your idea (with a fake door)

Why it matters: Before you build the product, build the page that sells it. If nobody clicks "Sign up," you saved yourself a month. We covered the methodology in How to Validate a Startup Idea in 48 Hours.

What you ship: A polished one-pager with email capture and a simple analytics setup so you can measure who's actually interested.

Starter prompt:
"Build a single-page landing page in Next.js. Headline, subhead, three benefit blocks with icons, social-proof strip, pricing teaser, email signup form, FAQ. Tailwind, dark theme. Email signups go to Beehiiv via API. Plausible analytics. Deploy to Vercel."

Ship time: 3–5 hours.

6

A "boring API wrapper" SaaS

Why it matters: The most underrated micro-SaaS pattern: take a public API that's hard to use and put a clean UI on top of it. Founders pay $19–$49/month for "I don't have to read the docs."

What you ship: A small SaaS wrapping one useful API. Examples: USPS address validation, OpenWeather batch lookups, image resize/optimize for ecommerce, transcript fetching from YouTube.

Starter prompt:
"Build a SaaS that wraps the [API name] API. Users sign up with email + Stripe. They get an API key. Endpoint: POST /process with their input. Charge per call or monthly. Dashboard shows usage and billing. Next.js + Supabase + Stripe. Deploy to Vercel."

Ship time: 10–14 hours.

7

A "show me the data" dashboard for a tiny audience

Why it matters: One indie dev built a niche analytics dashboard with Claude Code in 36 hours and had 15 paying users at $19/month within a week. Niche dashboards are the cleanest path from idea to MRR.

What you ship: A dashboard that pulls data from one source (Twitter, Stripe, GitHub, Spotify, your gym's API, whatever) and shows it in a way that's better than the source's own dashboard.

Starter prompt:
"Build a Stripe dashboard for indie founders. They paste a Stripe restricted API key. Show MRR, churn, top customers, MoM growth chart, and a 'this week vs last week' summary. Use Recharts. Daily email digest sent via Resend. Auth via magic link. Deploy to Vercel."

Ship time: 8–12 hours.

How to Actually Start (Not Browse)

Reading lists is procrastination. Here's the smallest possible action:

  1. 01Pick the project that made you feel something just now. Not the safest. The one that pulled.
  2. 02Open your terminal. mkdir weekend-project && cd weekend-project && claude
  3. 03Paste the starter prompt. Hit enter. Watch it scaffold.
  4. 04Iterate in plain English until it's shippable. Stop when "good enough" hits.

None of these the right idea?

Browse 100+ research-backed startup ideas with revenue data, market sizing, and Claude-Code-ready build prompts.

Browse Startup Ideas

TL;DR

  • Personal CRM, niche newsletter, Telegram bot, internal tool, fake-door landing page, API-wrapper SaaS, niche dashboard.
  • Each one fits in a weekend.
  • Each comes with a paste-ready prompt.
  • Pick one. Don't pick three.
1:1 Setup Service

Want help shipping one of these this weekend?

Book a 1:1 with John. Pick any project on this list—we'll get Claude Code installed, scaffold it together on a screen share, and have a working version live before the call ends. For founders who'd rather have a guide than a tutorial.

Book a Claude Code Setup Session (opens in new tab)

Stop browsing.
Open the terminal.

A weekend is enough. The rest is just typing.

Browse Startup Ideas