Genrupt: Amazon Creative Ops Platform for Sellers
Helped Genrupt become a paid Amazon seller platform with billing, team accounts, and reliable long-running creative workflows.
Project overview
Genrupt: Amazon Creative Ops Platform for Sellers overview
A short overview of the problem, solution, and result.
Starting Point
Genrupt had AI image and video generation working, but the product was not ready for serious customer usage. Paying seller teams needed subscriptions, organization billing, safer credit accounting, and long-running jobs they could trust.What Shipped
Subscriptions, organization billing, and credit-ledger accounting for paid seller teamsWhy It Mattered
Genrupt became commercially operable, with subscriptions, organization billing, credit-ledger accounting, webhook reconciliation, and safer cost controls for paying seller teams.Summary
Genrupt is an AI creative ops platform for Amazon sellers. It had a working AI media MVP, but paid seller teams needed more than generation screens.
The work moved the product into a production platform with subscriptions, credit billing, background jobs, Amazon seller workflows, and a safer way for agents to use the product.
Context
Genrupt was expanding from AI image and video generation into Amazon seller workflows: market analysis, review scraping, variations, listing builder, A+ content, and experimentation.
That expansion made the product more valuable, but it also exposed the limits of the MVP. Paid seller teams needed subscriptions, organization billing, credit accounting, reliable background work, and clearer recovery paths when long-running AI jobs failed.
As the surface grew, normal app screens were no longer the only interface that mattered. The product also needed a way for external agents and internal chat to run operations safely.
Approach
Move from demo-stage generation to seller-team operations.
The work focused on the product foundations that make a tool billable: organizations, subscriptions, credits, safer job handling, and cleaner workflow ownership.
Treat agents as a real product surface.
Instead of exposing the whole app implicitly, the agent surface used allowed capabilities, resource links, structured tool results, and recoverable long-running operations.
Solution
The product moved beyond a demo-stage generator into a more operable seller workflow platform with project, account, billing, and workflow surfaces for paid seller teams.
I helped ship the commercial foundation around subscriptions, organization billing, credits, and webhook reconciliation, then connected that foundation to long-running AI and Amazon seller workflows.
The Work
Turned billing and credits into a commercial foundation.
I helped add subscriptions, organization billing, credit-ledger accounting, webhook reconciliation, and safer cost controls so paid seller teams could use the product without the business relying on manual cleanup.
Made long-running AI workflows easier to trust.
Image, video, storyboard, and seller-workflow jobs needed to survive retries and give users clearer progress. Operation IDs, status and result tools, resource links, structured errors, idempotency keys, and additive progress notifications made those flows more dependable.
Opened the product to external agents safely.
The remote MCP and Agent Access surface let paid users connect external agents, discover allowed capabilities, and run market analysis, Amazon review scraping, listing-builder, and A+ workflows through a small permissioned tool catalog.
Impact
Genrupt became ready for paid seller teams because the work connected commercial infrastructure, AI workflow reliability, and agent access into one product foundation. That foundation helped support the first 100 paying customers.
Commercial foundation
Seller teams could become paying accounts.
Subscriptions, organization billing, credit-ledger accounting, and webhook reconciliation gave the product a cleaner path to paid usage.
Agent access
External agents could run product workflows safely.
The MCP and Agent Access surface exposed a small permissioned tool catalog instead of asking users to understand internal routes or job mechanics.
Workflow reliability
Long-running AI work became easier to trust.
Operation IDs, status tools, result tools, structured errors, and idempotency keys made generation and seller workflows more recoverable.
Product velocity
New seller workflows became easier to add.
Cleaner route, service, state, and request ownership reduced the fragility that had built up around earlier demo-stage features.
Visuals
The product proof is the shift from isolated generation into seller-facing workflows with the surrounding systems needed for paid usage.
