Cin7 Core Toolbox — Do amazing things that you cannot do in Cin7 Core

Import product/BOM data once via CSV, keep multiple Cin7 Core instances (and, later, other clients) in sync from a single canonical source.

Origin

Built for Casa das Natas (see docs/), designed to be reused across clients — org-scoped from the start.

Project notes

docs/PROJECT-NOTES.md is retired — it no longer holds the feature inventory or product reasoning. Current product/architecture/decision knowledge lives in the Spark Knowledge Obsidian vault instead, at 02 Products/Cin7 Core Toolbox — start at Cin7 Core Toolbox.md (the product map) or Cin7 Core Toolbox Current State.md (the verified current baseline). CLAUDE.md's own "Spark Knowledge" section explains how a session reaches it and how to load it selectively (/spark-context, /spark-preflight, /record-decision, /spark-close) — read that before starting work in a new session, especially on a different machine (Claude Code's own session memory is per-device and doesn't travel with the repo).

Stack

Next.js (App Router) + TypeScript on Vercel, Supabase (Postgres) for the canonical store, Vercel Cron for scheduled sync.

Scope (current phase)

1. CSV import of products (InventoryList template), Assembly BOMs and Production BOMs into the canonical schema.
2. Settings UI to connect multiple Cin7 Core instances per organization.
3. Sync engine: idempotent upsert (create/update) of products + both BOM types into every connected instance.
4. Reporting consolidator — later phase, after the feeder is working.

Data model

See supabase/migrations/0001_canonical_schema.sql. Key tables: organizations, cin7_instances, products, price_tiers, assembly_bom_lines (flat BOM), production_bom_versions / production_bom_operations / production_bom_items (routed BOM with work centres and resources), sync_state (per-instance sku → cin7_id map), import_batches / import_rows (CSV staging).

content_hash on products and production_bom_versions is trigger-maintained; the sync engine only pushes when it differs from sync_state.synced_hash for a given instance.

Reference CSV templates

docs/cin7-templates/ holds the exact Cin7 Core export templates the import parser targets: InventoryList (products), AssemblyBOM, ProductionBOM.

Cin7 Core API

Reference: https://dearinventory.docs.apiary.io/. See docs/cin7-api-findings.md for verified auth scheme, endpoints, pagination and rate limits — including that Production BOM push is supported via API (ProductionBom resource), correcting the original proposal's assumption that only Assembly BOMs would be needed. A few fields (Production BOM payload shape, Category/UOM write support) are still unverified against a live sandbox — confirm those before extending src/cin7/.

Setup

cp .env.example .env.local   # fill in SUPABASE_SERVICE_ROLE_KEY and ENCRYPTION_KEY
npm install
npm run dev

Supabase project: cin7core-feeder (pnzwjqjovxxdikxtfngq, Spark org, eu-west-1).