01
Normalize inputs
Read a supplier catalog CSV and map names, identifiers, price assumptions, and fees into a consistent research view.
OPZEIT LABS / INTERACTIVE PROTOTYPE
From supplier data
to clearer decisions.
Explore an early-stage sourcing research workflow built around catalog normalization, transparent unit economics, and human review. No marketplace data is claimed or retrieved.
TRANSPARENT ESTIMATESNO THIRD-PARTY API KEYSLOCAL CSV PROCESSINGHUMAN-APPROVED DECISIONS
01 / WORKSPACE
Sourcing analysis lab.
Import your own supplier CSV or explore synthetic sample data. All computations run in your browser.
EARLY PROTOTYPE
i Sample products and numbers are fictional. Nothing here verifies live prices, brand authorization, or Amazon selling eligibility.
CATALOG RECORDS 01
—
Products loaded
REVIEW CANDIDATES 02
—
Estimated thresholds met*
MARGIN / COST FLAGS 03
—
Require closer inspection
IDENTITY GAPS 04
—
Missing UPC and/or ASIN
CATALOG REVIEW
Product research queue
Select any product to inspect its estimated economics and review flags.
| PRODUCT / SOURCE | SELL PRICE | UNIT COST | EST. PROFIT | EST. MARGIN | RESEARCH STATUS |
|---|
* A “review candidate” only meets your currently selected estimated unit-economics thresholds. It is not an approval to sell, product match confirmation, or a buying recommendation.
02 / METHODOLOGY
A reproducible first-pass workflow.
Designed to support the research process—not to replace supplier verification, marketplace checks or human judgment.
02
Calculate transparent estimates
Compute illustrative unit profit, margin and ROI using explicit, editable assumptions. No hidden score or invented sales velocity.
03
Flag for human review
Highlight identity gaps, weak margins and missing cost information. Verify ASIN match, listing permissions and IP risks separately.
04
Export the research queue
Download your filtered analysis as CSV, then move verified candidates into a separately controlled sourcing workflow.
03 / APPLIED AI ROADMAP
Claude in development. API next.
Claude has been an active development assistant across the founder’s software projects, supporting architecture, code iteration, debugging and testing. This particular prototype runs locally in the browser with JavaScript; it does not currently call Claude API, Amazon services or live marketplace feeds.
PLANNED / NOT INTEGRATED
Supplier document understanding
Investigate Claude for extracting structured fields from unstructured vendor PDFs and spreadsheets with validation checks.
Research explanations
Explore clear, auditable summaries of missing fields, price assumptions, and issues requiring manual follow-up.
Engineering productivity
Continue Claude-assisted design, code review, debugging, tests, documentation and iteration on internal sourcing tools.
◉
Human decisions are a hard boundary.
Any future AI functionality will be evaluated for reliability, privacy, accuracy and compliance. No autonomous purchasing, account access, financial advice or trading execution is part of this demo.