Copiara
In development. Four surfaces over one catalog, price, and order model, with the order backbone, the AI layer, and enterprise buying already running in production.
Role
Studio · build + ship
Status
In development
Year
2026
Surfaces
Web
Live
In build
The problem
Copiara is the operating system for wholesale buying. A distributor gets a buyer portal where their customers self-serve, a rep console where inside sales works the book, an admin console where the business configures catalog and pricing, and a platform console for the operator, all reading one shared catalog, price logic, and order record. Product data lives and improves inside the platform: imports enrich rather than overwrite, every change is kept as history, and duplicate parts get caught before a buyer ever sees them. The AI concierge sits on the storefront and answers from the same permissions the buyer already has, so it cannot quote a price or a stock position that account is not entitled to. When it wants to act, it proposes and a human confirms.
What we built
- Four surfaces over one data layer: buyer portal, rep console, admin console, platform console
- Order backbone: shipments, inventory reservations, honest backorders, returns, credit memos, invoices
- Product data that improves in place, imports enrich instead of overwriting, with full change history
- An AI concierge scoped to the buyer’s own permissions, staging actions for a human to confirm
- Enterprise buying: punchout, PO and GL coding, approval chains, budgets, three-way match
- Buyer company hierarchies with shared credit pools and national accounts
- Readable audit trail with per-tenant hash chains, a permission matrix, and tenant data export
Stack
- Frontend
- Next.js 16React 19Shopify PolarisTypeScript
- Backend
- FastAPIPython 3.12Alembic migrations
- AI / ML
- Vertex AIGemini concierge + rep copilotServer-side tool runtime
- Data
- PostgreSQL (multi-tenant)Product revisions + provenance
- Infra
- Cloud RunVercelSSO, SAML, SCIM
What’s next
Turn on subscription billing, land the first distributor pilots, and prove punchout against a live enterprise buyer.
Have something that looks like this?
Tell us what is broken. We’ll tell you what the first week looks like.
Next case study →
Smile PreVue