1. Manufacturer discovery
Search engines, industry directories, trade fairs and international B2B profiles may help us discover companies. Those sources do not automatically become the product record.
2. Official website verification
We compare the company name, contact information, address, product scope and About information. Uncertain domains remain in a manual review queue.
3. Source collection
Product data is collected primarily from official websites, product pages, catalogues and datasheets. The collector respects robots.txt, rate limits and access restrictions. It does not bypass login or CAPTCHA controls.
4. Extraction and normalization
Rules and optional AI convert public source text into structured claims. AI is not permitted to fill missing values. Units and terminology are normalized while the original value and exact source text are retained.
5. Review and maintenance
New records remain drafts until approved. A missing page is marked as source missing, not discontinued. Discontinued status requires an explicit manufacturer statement.
Database inclusion and Export Readiness grades are research aids, not endorsements, certifications or guarantees.