We build software that
understands garments.
Merchant Pro was built by people who spent years watching merchandisers copy data between spreadsheets. We believed AI could solve this — and it can.
The problem we saw
In every RMG factory we observed, the merchandising workflow looked the same: a buyer sends a tech pack PDF, a merchandiser opens Excel, and the next two days are spent copying data — RM codes, colorway names, material descriptions — from the PDF into a BOM, then from the BOM into a tracker, then from the tracker into an email.
Every transfer introduced the possibility of error. A wrong consumption value would cascade through 17 RM items and 11 colorways — 187 wrong cells — before anyone noticed. And by then, the cost sheet was already wrong, the booking was already done, and the ex-factory date was already at risk.
We built Merchant Pro to automate every mechanical step in that workflow, so merchandising teams spend their time on judgment calls — pricing, supplier selection, quality checks — not on copy-paste.
Our philosophy
We believe the right model for AI in manufacturing is human-in-the-loop, not AI replacing humans. Claude Vision reads the tech pack and extracts what it can. Your team reviews it, fills the gaps, and confirms the data. The AI learns from every correction.
This means the system gets smarter with every tech pack parsed for your buyer. Reference files from previous seasons, buyer standards, and standard RM code lists all feed into the context — accuracy improves continuously.
We also believe in complete traceability. Every number in your RM tracker should trace back to a source document. Every change should be attributed to a user with a timestamp. That is how factories earn buyer trust.
Built on infrastructure that factories can trust.
Claude Vision (Anthropic)
Anthropic's Claude Sonnet Vision is the most capable document-understanding AI available. It reads free-form PDF tech packs without templates, understands industry terminology, and extracts structured data at human-level accuracy.
Next.js + Prisma + PostgreSQL
Merchant Pro is built on Next.js 14 with App Router for a fast, server-rendered UI. Prisma ORM on PostgreSQL provides type-safe, relational data storage with full multi-tenant isolation. Your organisation's data never mixes with another's.
Cloudflare R2 + Redis
Generated Excel documents and uploaded PDFs are stored in Cloudflare R2 — S3-compatible object storage with no egress fees. Redis powers async job queues for long-running parse operations, keeping the UI responsive during heavy processing.
How we build.
Human-in-the-loop
AI suggests, humans decide. Every extraction is reviewable and correctable. We will never ship a system that overwrites human judgment without confirmation. The merchandiser is always in control.
Data integrity
All calculations are traceable. Every number in the RM tracker links back to a confirmed source — a consumption value in the BOM, a unit price in the cost sheet, an order quantity from the PO. You can always answer 'where did this number come from?'
Industry specificity
We built for RMG, not generic supply chain. The system understands RM codes, colorways, tech packs, BOM structures, and ex-factory logistics natively. We speak your language because we learned it from the factory floor.
Continuous improvement
Human corrections improve future parses. Reference files from previous seasons build buyer-specific accuracy. Every tech pack your team processes makes the next one faster and more accurate — the platform compounds.
We built this for your team.
Start with one tech pack. See the difference in 60 seconds.
Sign in to Merchant Pro