Clinical Trials Scraper — ClinicalTrials.gov, Phases & Sites

Search ClinicalTrials.gov and get one row per trial: NCT id, recruitment status, phase, enrollment, sponsor, conditions, interventions, eligibility and the countries with sites.

No API key, no account. $0.01 per trial — a search that matches nothing is recorded and not charged.

Input

{ "query": "melanoma", "status": "RECRUITING", "phase": "PHASE3", "maxTrials": 100 }

Bulk, restricted to a country:

{
  "queries": ["semaglutide", "tirzepatide"],
  "status": "RECRUITING",
  "country": "Germany",
  "maxTrials": 250
}

Field	Type	Notes
query	string	Condition, intervention, sponsor or free text — melanoma, semaglutide, Pfizer.
queries	string[]	Up to 10 searches per run.
maxTrials	integer	1–500 per term. You are charged per trial returned.
status	enum	ANY, RECRUITING, COMPLETED, TERMINATED, …
phase	enum	ANY, EARLY_PHASE1, PHASE1–PHASE4.
country	string	Country or city with a trial site. Empty = worldwide.

Output

{
  "ok": true,
  "nctId": "NCT00920907",
  "url": "https://clinicaltrials.gov/study/NCT00920907",
  "title": "Comparison of Ipilimumab Manufactured by 2 Different Processes…",
  "status": "COMPLETED",
  "phases": ["PHASE1"],
  "studyType": "INTERVENTIONAL",
  "enrollment": 99,
  "conditions": ["Advanced Melanoma"],
  "interventions": ["Ipilimumab"],
  "sponsor": "Bristol-Myers Squibb",
  "startDateRaw": "2009-08",
  "startDate": null,
  "locationCount": 7,
  "countries": ["United States"],
  "eligibilitySex": "ALL",
  "minimumAge": "18 Years",
  "briefSummary": "The purpose of this clinical research study is to…"
}

What it gets right

• ⭐ Partial dates are not invented. ClinicalTrials.gov returns "2009-08" when the sponsor only ever committed to a month. That is real information. The raw value is kept in startDateRaw, and startDate stays null rather than silently becoming the 1st — so a date you can compute with is always a date somebody actually stated.
• Filters are really applied, server-side. status: "COMPLETED" means every row is completed, not a full result set you then have to filter yourself and pay for.
• It pages past the API's 100-per-request ceiling via nextPageToken, stops when the token stops advancing, and is bounded so it cannot spin.
• Nested modules are flattened into one flat row — sponsors, conditions, interventions, eligibility and site countries all sit at the top level, ready for a spreadsheet.
• countries is de-duplicated from the site list, so a 40-site trial gives you 3 countries, not 40 repeats.

Pricing

$0.01 per trial returned. No start fee. A search matching nothing produces a trialCount: 0 record and is not charged.

Limits

• 10 search terms per run, up to 500 trials each.
• Returns the protocol record: status, design, sponsor, eligibility, locations. Posted results data (outcome measures, adverse events) is not included.
• country matches trial sites, not the sponsor's country.
• Data is US-government public domain, updated by sponsors — lastUpdateRaw tells you when.

FAQ

Do I need an API key? No.

Why is startDate null when startDateRaw has a value? The source gave a month, not a day. See above — we do not guess.

Can I get outcome results? Not in this Actor; it returns the protocol section.

Changelog

• 0.1 (2026-09-10) — first release. One row per trial, token paging, server-side status/phase/ location filters, partial dates preserved.