Machine Shop / Contract Manufacturer Capability Detector

What you are looking at

A list of job shops gives you a name, a headcount and a zip code.

Why it costs you

It does not say which shops run 5-axis, hold AS9100, or cut Inconel. So reps qualify on the call instead of before it, territories get planned on guesswork, and a chunk of the book is shops that were never going to buy. That data exists only as prose on a capabilities page, so no database holds it.

What this actor does

Point this actor at a list of shop websites. It crawls each shop's own capability, equipment and quality pages and returns machine brands, certifications (AS9100, ISO 9001, IATF 16949, ISO 13485, ITAR, NADCAP, AWS, AISC), processes from 5-axis and Swiss through welding, fabrication, foundry and finishing, materials, industries served, the advertised tolerance, and the shop-floor ERP or quoting system where one is exposed.

Start here

Paste a handful of shop domains into startUrls and run it. Compare the output against shops you already know before you scale it to a territory.

Output fields

• vertical_relevance
• company
• website
• final_url
• top_detection
• vendor
• confidence
• evidence
• machine_brands
• machine_brand_count
• certifications
• processes
• materials
• industries_served
• tolerance_claim
• has_rfq_form
• all_detections
• categories
• pages_crawled
• status

Measured accuracy

96 to 100% on qualified shops across 8 stress rounds on fresh OpenStreetMap samples in 5 US regions. Processes fill 85-100%, materials 54-90%, certifications 23-50%. Machine brands are the sparsest field (12-23%) because many shops do not publish an equipment list.

Deploy

$APIFY_PUSH_TOKEN=$APIFY_TOKEN_PERSONAL python3 push_to_apify.py

SDK-free by design: plain Apify REST via run env vars. Do not reintroduce the apify/crawlee SDK without pinning every transitive dependency.