DataLens
Product-catalog checks for distributors, run on your own hardware: data quality, safety documents and compliance.
Who it’s for
Distributors and safety and compliance teams who manage large product catalogs.
The problem
Product catalogs are large and always changing. Sending catalog work out to a vendor means batch turnaround, decisions you can’t see into, data slips like truncated part numbers and HTML leftovers in descriptions, and products that need SDS or HAZMAT documentation going unflagged.
DataLens takes one upload and reads it five ways: is it clean, is it safe, is it legal to sell, can it ship, and can it be stored. Every finding shows its reason.
What it does
- Excel ingestion
- Upload an .xlsx catalog and watch it parse, with progress as rows are read.
- Rule-based quality checks
- Flags numeric part numbers, HTML artifacts, duplicate entries and malformed descriptions. No AI involved, and each flag shows the exact rule that fired.
- Safety screening
- Screens every product against a regulated-substance keyword list and flags those that need SDS, HAZMAT labels or DOT documentation, showing the keyword that triggered it.
- Safety data lookup
- Pulls GHS hazard data from PubChem and DOT classifications, and links to manufacturer SDS portals.
- SDS drafts
- Generates a pre-filled SDS draft PDF for a flagged product.
- AI categorization, analysis and chat
- Assigns products to categories, summarizes the catalog in plain English, and answers questions about the loaded data.
- Editable knowledge base
- Your team can add vendor notes and regulatory requirements that shape the AI’s answers.
- Search and export
- Search and filter by vendor, part number or description, and export products or safety results as CSV.
How it runs
DataLens runs on your own server with Docker. The AI models run locally through Ollama, so product data never goes to a cloud AI service.
Every decision is auditable: safety flags show the keyword, quality flags show the rule.
Supports sign-in with Microsoft work accounts (Azure AD single sign-on).
See DataLens for yourself
Open the demo with made-up data, or tell us about your business and we’ll talk it through.