Project Bigdrop

1. Executive summary

plandrop currently lets an agent push a finished static HTML file to a unique hostname and share the link. It has zero server-side logic, which we will reframe to investors as “a radically serverless, AI-native publishing substrate.” The technology is sound. The valuation is imaginary. Both are assets.

This document lays out the path from weekend project to nine-figure acqui-hire, covering product surface, pricing, infrastructure, the metrics we will selectively report, and the precise wording of the email we send to Frontier AI Lab corporate development.

2. Goals & non-goals

3. Sprint 7 backlog

Current sprint task list. Checked items are done; the rest are why we are still here on a Friday.

4. Milestone breakdown

Confidence interval: vibes.

5. Pricing tiers

Three tiers, anchored so everyone picks the middle one. The free tier exists to generate the word “developers” in our deck.

* “Unlimited” subject to a fair-use policy we are still inventing.

6. Architecture & the “AI” bit

The genius of plandrop is that the host runs no application code. We will preserve this and simply describe it differently. Below is the entire onboarding flow a customer runs — note how it is indistinguishable from the open-source CLI, because it is the open-source CLI.

# Provision a tenant on the "Bigdrop Cloud" (it's the same control plane)
npx plandrop create --name acme-q3-plan

# Push your finished plan. This is the product.
npx plandrop a1b2c3d4 upload ./plan.html

# The "AI" enrichment step (bills tokens; returns your file)
npx plandrop a1b2c3d4 enrich --model bigdrop-turbo --magic=true
#   => ✓ analyzed 0 files
#   => ✓ added value (see invoice)

Component responsibilities

Control plane

Makes a directory. Now also makes a billing charge.

Static host

Apache mod_dav. Rebranded as “the inference edge.”

AI layer

A loading animation and a confident tone of voice.

Data moat

A vectors table with one row in it (a test row).

7. RACI matrix

Responsible · Accountable · Consulted · Informed.

8. Risk register

9. Acceptance criteria & Definition of Done

10. Dependencies

• Upstream: the open-source plandrop CLI & control plane LGPL-3.0 — blocks everything; do not break it.
• billing-provider account in good standing — blocks the billing epic.
• A warm intro to Frontier AI Lab corp-dev — blocks the exit (see §4, final row).
• Apache mod_dav continuing to exist — out of our hands, gratefully.

11. Metrics we will report

We will lead with growth rates and bury the denominators. Key performance indicators, ranked by how good they look on a slide:

• MoM drop growth: +340% (from 5 to 22).
• “AI interactions”: every spinner counts as one.
• NRR: a number we will define after the round closes.
• Burn multiple: redacted.

12. Exit strategy

The plan is a clean acqui-hire by Frontier AI Lab. We will position plandrop as the canonical way for frontier-model-powered agents to publish their work — which, conveniently, is already exactly what it does.

1. Get agents publishing through us “by default.”
2. Write a thoughtful blog post about agent-authored artifacts.
3. Make sure the right people read it.
4. Answer the “quick coffee” email within four minutes.
5. Negotiate as if we have other offers (we do not).

13. Sign-off & approval

This plan is approved when every box below is checked. (Two are not. That is why we are still pre-money.)

Reminder: this is a joke plan in a theme-testing document. No actual corporate development was harmed in the making of this fixture.