Getting started

Discovery vs enrichment

StartupHub is built around proactive discovery: we find companies from domain registrations, government registries, GitHub orgs, and live websites before most of them appear in press. Generic web-search APIs (Tavily, Linkup, Parallel, Exa) answer a different question: what is being written about right now. Both are useful. They are not interchangeable.

How we discover new startups

Discovery is a set of scheduled ingestion pipelines. Each pipeline watches a primary signal, stages candidates, probes the website, fingerprints the stack, and promotes a row into startups. No news article required.

After promotion, enrichment fills sectors, description, founders, funding, and tech_fingerprint. Fresh promotes may have empty sectors for a few hours; use created_at for "added this week" and founded_date for "founded this month" (domain/reg filing date until enrichment overwrites with grounded data).

StartupHub vs Tavily / Linkup / Parallel

Recipe queries

These are the queries buyers actually run. All use the same endpoint: GET /startups.

New cybersecurity startups in Israel (this month)

curl -G "https://www.startuphub.ai/api/v1/startups" \
  -H "Authorization: Bearer sk_live_..." \
  --data-urlencode "country=Israel" \
  --data-urlencode "sector=Cybersecurity" \
  --data-urlencode "founded_after=2026-07-01" \
  --data-urlencode "sort=created_at.desc" \
  --data-urlencode "limit=50"

New AI startups in Canada (added since June)

curl -G "https://www.startuphub.ai/api/v1/startups" \
  -H "Authorization: Bearer sk_live_..." \
  --data-urlencode "country=Canada" \
  --data-urlencode "sector=Artificial Intelligence" \
  --data-urlencode "sort=created_at.desc" \
  --data-urlencode "founded_after=2026-06-01" \
  --data-urlencode "limit=100"

Stealth AI on .ai / .io with verified founder emails

Requires Pro Lite+ for stealth=true and has_known_emails=true.

curl -G "https://www.startuphub.ai/api/v1/startups" \
  -H "Authorization: Bearer sk_live_..." \
  --data-urlencode "stealth=true" \
  --data-urlencode "domain_tld=ai,io" \
  --data-urlencode "has_known_emails=true" \
  --data-urlencode "sort=created_at.desc"

What POST /enrich does

Enrichment is for a known company (name + ideally website). It does not discover net-new companies from an open-ended prompt. When the pipeline runs, every paid call enables live Google Search grounding (same engine as admin Fetch Info), then:

1. Gemini grounded extraction (description, sectors, HQ, funding with citation guards)
2. Direct website scrape (socials, logos, page text)
3. Tech fingerprint (CDN, hosting, analytics, payments, ATS, …)
4. LinkedIn employee count, open roles sync, GitHub commit emails where available
5. Founder/team minting only when grounding sources verify the person matches the domain

curl -X POST "https://www.startuphub.ai/api/v1/enrich" \
  -H "Authorization: Bearer sk_live_..." \
  -H "Content-Type: application/json" \
  -d '{"name":"Acme Security","website":"https://acme.security","force":true}'

Flags: fast:true returns cache only;force:true re-runs even if enriched within 7 days. See Enrich Company.

Recommended agent workflow

1. Sourcing: poll GET /startups?sort=created_at.desc with your ICP filters (country, sector, founded window).
2. Deep dive: for rows you care about, call POST /enrich or fetch GET /startups/:slug.
3. Buzz / narrative: use GET /news/mentions or a web-search tool for "who got written about this week" (orthogonal to discovery).
4. Lists: POST /lists/ai-generate combines grounded search with directory matching when you have a natural-language brief.

Product overview: New Startups API. Full reference: API docs.