Revenue Intake Ledger

A DynamoDB-backed revenue follow-up room for small AI builders.

Comment

Inspiration

AI agents are making it easier for one person or a small team to launch many products, submissions, videos, stores, and experiments. That is exciting, but it creates a new problem: the work after creation gets messy.

Proof, deadlines, cloud credits, cost checks, payout tasks, sponsor requirements, and human approvals spread across tabs and tools. Revenue can be lost not because the product was bad, but because the follow-up was not organized.

Revenue Intake Ledger is my answer to that future problem. It turns an AI-run portfolio into a durable operating ledger for money.

What it does

Revenue Intake Ledger is a Vercel app backed by AWS DynamoDB. It tracks revenue opportunities, submitted lanes, evidence items, payout tasks, award dates, risks, AI-suggested next actions, and status history in one operating room.

The H0 lane itself is stored as live DynamoDB data, so the product uses its own database-backed ledger to manage this submission.

How I built it

The frontend is a Next.js app deployed on Vercel. API routes read from DynamoDB in production through encrypted Vercel environment variables.

DynamoDB uses a small single-table shape:

  • OPPORTUNITY#<id> / PROFILE for each revenue lane
  • OPPORTUNITY#<id> / EVIDENCE#<id> for proof records
  • OPPORTUNITY#<id> / PAYOUT#<id> for follow-up tasks
  • OPPORTUNITY#<id> / STATUS#<timestamp> for status history
  • WORK_QUEUE#open / DUE#<date>#OPPORTUNITY#<id>#PAYOUT#<id> for open follow-up actions

The table is RevenueIntakeLedger in us-east-1, on-demand billing, and server-side encryption is enabled. Runtime access uses a dedicated IAM user with a DynamoDB-only least-privilege policy. The AWS credit is applied, but the setup stays deliberately tiny because promotional credits are useful but not a hard spending cap.

AWS Database Used

Amazon DynamoDB is the source of truth for the production app.

Public API proof:

  • /api/health reports database=dynamodb
  • /api/opportunities reports source=dynamodb:RevenueIntakeLedger
  • /api/h0-bundle loads the H0 packet with PK = OPPORTUNITY#h0
  • /api/action-queue loads open result, payout, and AWS-usage follow-up tasks with PK = WORK_QUEUE#open
  • /api/evidence includes the updated demo video and DynamoDB proof
  • /api/payout-tasks stores payout follow-up tasks
  • /api/proof returns h0Ready=true and live DynamoDB metrics

Challenges

The biggest challenge was keeping the submission honest. A Vercel URL and a nice UI are not enough for H0. I added a proof board, cost boundary, DynamoDB readback, AWS storage screenshot, public proof APIs, and a final submitted-state stopline so the app does not overclaim.

I also avoided widening runtime permissions. A direct table-structure update through the runtime IAM user was denied, which is good: the deployed app can read and write its operating data, but it cannot administer the database.

Accomplishments that I am proud of

  • Live Vercel app connected to DynamoDB
  • DynamoDB table seeded with operating records
  • Single-table query proof through /api/h0-bundle
  • Open work queue proof through /api/action-queue
  • Public proof API for database source and submission readiness
  • Status history stored as durable records
  • Redacted AWS storage proof screenshot captured
  • Under-3-minute functioning demo uploaded
  • Cost-aware AWS setup using the provided promotional credit route

What I learned

The best database-backed product is not just a CRUD app. The database should carry the product memory. In the agent era, that memory becomes even more important: it keeps humans and AI aligned after the first launch.

Here, DynamoDB is the durable memory for revenue follow-up, evidence, status changes, open action queues, and payment tasks.

What's next

Add authenticated team workspaces, append-only status updates from the UI, Google Calendar reminders for award dates, payout document tracking, and AI-generated follow-up summaries for founders and small teams.

Links

Live app: https://revenue-intake-ledger-public.vercel.app/ GitHub: https://github.com/daideguchi/revenue-intake-ledger Demo video: https://youtu.be/sGeT3OtkBks Public proof API: https://revenue-intake-ledger-public.vercel.app/api/proof H0 bundle API: https://revenue-intake-ledger-public.vercel.app/api/h0-bundle Action queue API: https://revenue-intake-ledger-public.vercel.app/api/action-queue

#H0Hackathon build note

I also published a H0 build note before the June 29 submission deadline: How Revenue Intake Ledger uses AWS Databases after AI shipping. It explains who the product helps, the follow-up problem it solves, how AI agents are used, and why DynamoDB is the product memory instead of a checkbox integration.

Built With

Share this project:

Updates