TradeIndia Suppliers Scraper — B2B Directory

Search TradeIndia's B2B supplier directory by product/service category and get every manufacturer listed — company name, location, business type, product detail, and (optionally) full profile enrichment: address, GST number, owner name, seller tier. No login, no cookies, no browser.

Why this actor's shape differs from a normal search scraper

TradeIndia's robots.txt closes every pagination mechanism this actor could otherwise use — query-string paging (?page=) is disallowed site-wide, and the internal REST/catalog APIs a prior recon session found are separately blocked (/restapi/, /catalog_apis/, /ai-search/). This actor does not try to route around that. Each category keyword returns one page's worth of suppliers (typically ~28) — an honest ceiling reported on CATEGORY_SUMMARY, not implied to be a partial slice of a larger reachable set.

A plain slugified guess only matches TradeIndia's own category taxonomy about 13% of the time — "solar panels" isn't solar-panels, it's rec-solar-panels; "cotton fabric" isn't cotton-fabric, it's raw-cotton-fabric. So a category term that doesn't match directly falls back to the site's own published sitemap (sitemap_manufacturers_index.xml — robots.txt-endorsed, ~110,000 real category slugs) and resolves to the best token-overlap match. This fallback loads once per run, lazily, only if at least one term needs it.

Why one actor, not "search" + "profile lookup" as separate actors

A category search naturally produces profile URLs, and a user who wants company details (address, GST, owner) inlined shouldn't have to glue two actor runs together. enrichProfiles is a checklist toggle on one pipeline; profileUrls (direct profile input, skipping search) reuses the same enrichment code path.

What you get

SUPPLIER

Field	Notes
companyName, city, state, country, businessType
establishedYear, memberSinceYears	Not the same thing — memberSinceYears is a duration ("11 years on TradeIndia"), not a year or date
categoryName, productName, productDescription, priceRange
profileUrl, productUrl
numberMasked	Always true for unauthenticated visitors — TradeIndia's own privacy gate, not something this actor bypasses. Present so you know not to expect a working phone field, rather than reading a silent None as a scrape failure.

With enrichProfiles: true (or via profileUrls), also: address, gstNumber, ownerName, designation, catalogDomain, intro, sellerTier.

CATEGORY_SUMMARY — one per category term

rowsOnPage (the honest per-category ceiling), profilesEnriched, finalUrl, and notes — including whether the sitemap fallback had to resolve your term to a different real slug.

ERROR — one per input that failed

Known limits

• ~28 suppliers per category term. This is TradeIndia's own per-page batch with no policy-compliant way to reach a second page (see above). Search several related keywords to broaden coverage instead of expecting deep pagination on one.
• Phone numbers are always masked for unauthenticated access — a TradeIndia privacy feature, not an extraction gap.
• This host rate-limits hard by IP. Recon burned through a local dev IP's budget after ~40-50 requests in a few minutes; TLS-profile rotation alone did not clear the block, a different exit IP did. Residential proxy is the default for exactly this reason — not a generic caution, load- bearing for this specific host.

Proxy

Residential is the default and matters more here than on most targets in this portfolio — see "Known limits" above.