OneProposal
ParkedA parked AI proposal product covering research, generation, refinement, hosted output and failure recovery.
I built OneProposal to turn a freelance project brief and background information into a proposal and a hosted page. The product is parked. The public sample preserves the output format.
What it actually is
I built a proposal-generation and refinement product with a front end and a separate generation engine. The work covered research, extraction, a cover letter, a hosted proposal page and checks on the generated output. It also included the contracts and callbacks connecting the two parts, plus recovery when a request failed.
The hard parts
One recurring problem was contract drift. The engine could return a field the front end did not use, leaving a feature partly implemented. I used explicit request and callback contracts and regression checks to make that boundary easier to verify.
Owner tests exercised the payment and generation paths. Customer validation required a separate acquisition step. The acquisition sprint ended without reaching its goal, and I parked the project.
- Updated this page to reflect the parked product. The public sample shows the output design, not a new generation or a customer result.
- Parked the product before reaching the acquisition goal. Completed checkouts were owner tests, not external customer validation.
- Moved the strongest model to the default for drafting and kept the cheaper one for refine and follow-up, so quality goes where it matters and cost stays in check.
- Ran an audit and worked through the findings. Owner tests exercised the payment and generation paths. Added refinement, follow-up generation and richer research output.
- Quiet launch to friends and family. Flipped payments to live mode, moved onto its own domain, wired analytics and an in-app feedback widget, and shipped credit purchase plus a batch of readiness fixes in one go.
- Made live research a hard requirement, so a proposal has to be grounded in real signals about the client instead of invented ones. Shipped a cleaner editorial template and verified it end to end.
- Closed a productization sprint. Added a finished one-page proposal and regeneration from a different angle. Moved from the heavier agent framework to direct model calls.
- Started building the proposal engine.