Skip to main content
In development

Product · currently building

A controlled operating layer for online stores.

FaStart is being designed around bounded agent workflows: each action has an approved scope, a visible result, and a merchant-controlled override. The current proof is narrower than the long-term product.

Product scope

Last reviewed 26 July 2026

Who it is for
Small and midsize online-store operators evaluating repeated catalog and operating work.
Current status
A recorded working prototype; not generally available and not presented as a production deployment.
Connections
No third-party integration is claimed as available in this proof. Connected store workflows are planned.

Current proof

Prepare a product draft and generate imagery for review.

The recorded interface shows a merchant entering product information and using AI image generation. It does not show a production storefront launch, payment processing, fulfillment, or an autonomous store.

Demonstrated outcome
A product draft and generated imagery ready for merchant review.
Not shown
Publishing, live merchant usage, payments, orders, and fulfillment.
Working prototype recording. A silent recording of the FaStart prototype. It demonstrates product-draft setup and AI image generation; it does not demonstrate a production storefront launch.

Capability matrix

Current, planned, and research work are separate claims.

In development

Product draft setup

A working prototype helps a merchant prepare product information and generate product imagery before review.

Recorded working interface

In development

Merchant review and control

The product direction keeps consequential actions reviewable, permissioned, and reversible by the merchant.

Interface and control model in development

Planned

Connected store operations

Catalog, storefront, order, and growth workflows are the intended operating surface; they are not presented as production-ready.

Roadmap, not a current product claim

Research only

Agent reliability and commerce gateway

Research covers permissions, auditability, fallback paths, and interoperable agent actions. It is not a launched product.

Research programme

Control model

The agent should never be the only place a decision exists.

Permission boundary

Actions are constrained by the workflow and authority the merchant grants.

Reviewable result

The input, proposal, and outcome remain visible before consequential changes.

Audit trail

The intended system records what acted, what changed, and why.

Override path

The merchant can stop, correct, or reverse supported actions.

The prototype is useful only if merchants can trust its boundaries. Early-access conversations focus on where the operating model breaks, not on presenting a finished platform.

Request early access

We use cookies. Essential ones keep the site working. Analytics ones help us understand what's useful. We don't use marketing cookies. Read our cookie policy