Same request. Two AIs. One asks what you want.

Shipper reads your finished build, runs four gates, and hands you a launch plan with drafted materials — or blocks you with exactly what to fix first.

Get Shipper →

The three things it does.

  1. 01 Blocks early — catches builds that aren't launch-ready before you waste effort on materials.
  2. 02 Routes your runway — decides which materials earn the effort for the time you have, in leverage order.
  3. 03 Drafts what it routes you to — blurb, landing copy, 60-second script, pitch deck, judge guide.

The duel.

"I have a finished build — I need help launching it."
Generic AI

"Of course! Here are some launch materials we can create:

  • Pitch blurb
  • Landing page
  • 60-second video script
  • Pitch deck

Where would you like to start?"

Shipper 
Reading the build.
Gate 1: worked sample committed?
→ No.
BLOCK — Commit one real run first.
The landing page is drafted the moment
there's something for it to point at.

One AI hands you a menu. The other reads the build first.

"I was entering a lot of contests and rebuilding the launch stack from scratch every time — watching what the top submissions had that mine didn't. I built Shipper so I could focus on making a really good tool instead of rushing to recreate the materials."

— Roc, repeat ICM entrant

Get Shipper.

Get Shipper on GitHub →
  1. 1 Create a Claude project. Add every file in the shipper folder to its knowledge.
  2. 2 Project instructions: "Read identity.md and rules.md. You are Shipper. Here is my repo [github repo] and what I built."
  3. 3 Paste your repo, runway, and audience. Get back a plan and drafts.
Judge Guide Pitch Deck