Model your products once. Publish them everywhere — with a per-channel readiness score that shows exactly what each surface still needs before you ship. New attributes are data, not migrations.
No. 01 The schema
Products and variants attach to families of typed, localizable, channel-scopable attributes — text, measurement, money, select, richtext, media. The schema is the product.
Define a field in the UI or over the API. No migration, no deploy, no schema-change ticket waiting on engineering.
Every family compiles to a Zod validator generated from its own definition — so data is correct before it leaves your catalog.
Each mutation emits a typed event. A connector is just something that subscribes to it. Nothing is bolted on.
// generated from the attribute family — never hand-written export const Apparel = z.object({ title: z.record(Locale, z.string().min(1)), material: MaterialEnum, price: z.record(Market, Money), net_weight: Measurement.optional(), gallery: z.array(MediaRef), });
No. 02 Channel readiness
A merchandiser completes a product's attributes; a per-channel score shows exactly what each surface is missing — before you publish, not after a rejection.
Amazon wants bullets and A+ content; BigCommerce wants a shipping weight; Shopify wants a GTIN. Each score is computed against that channel's own rules.
Readiness validates from the same Zod schemas that back your data, so a surface can never ask for something your catalog can't describe.
When a channel reads 100%, publishing emits the events its connector already listens for. Your catalog stays the source of truth.
No. 03 Honest comparison
The PIM category went enterprise and never came back. minipim is the one the rest of us should have had all along — open source, an afternoon to stand up, free to start.
No. 04 Connectors
Connectors subscribe to the events your mutations emit. minipim pairs natively with Throttle, our open-source commerce engine — or with any storefront via the open API and event bus. Maintenance-heavy marketplaces are managed for you in Hosted.
No. 05 Enrichment · on the roadmap
AI enrichment is built in — it drafts copy, normalizes messy attributes, translates locales, and fills the exact gaps the readiness score flags. You review and accept.
Pricing
The whole engine is open source. Pay only when you want the managed connectors and auth.
The complete schema engine and core connectors, on your own infrastructure.
Everything in open source, fully managed — plus the connectors that are a pain to run yourself.
Clone from GitHub and self-host, or start free on Hosted and add the heavy connectors when you need them.