ClinicalTrials.gov Study Research — tool reference

Server URL: https://toolcargo.com/mcp/clinicaltrials · Scope: connector:clinicaltrials. Activate Site Audit free and connect with ToolCargo OAuth or a scoped API key. No source account, API key or paid provider. Each successful search page or exact record lookup uses one shared-plan call.

clinicaltrials_search_studies

{ "topic": "sleep apnea", "status": "RECRUITING", "limit": 5 }

Supply a plain human condition topic of 1–200 characters. Whitespace and Unicode normalization are disclosed. Boolean operators, source DSL, wildcards and NCT-ID queries are rejected. Source matching is not guaranteed literal, exhaustive or translated. Optional status uses a source overall-status enum. Limit is 1–10.

Pass nextCursor with unchanged topic, status and limit to request another explicit page, up to page 1000. The cursor also binds fixed selected fields and sort order. Pages are mutable; empty pages may still have continuation tokens, and a later total count may be omitted and returned as null. Source responses do not echo query identity; the cursor binding is ToolCargo’s control, not upstream query attestation.

clinicaltrials_study_record

{ "nctId": "NCT04368728" }

Use one canonical NCT ID with eight digits. Surrounding whitespace and case normalization are disclosed; NCT00000000 is rejected; source aliases and redirects are not followed. A returned record must have the exact canonical identity. not_found requires the recognized NCT-specific source 404 text, and means only that this endpoint returned absence at observation time. Generic errors never become absence.

Study status and dates

record.overallStatus and record.hasResults are independent. A completed study can have no posted results. Availability does not establish outcomes, efficacy, safety or completeness. Dates retain the source’s month or day precision without invented days. startDate, primaryCompletionDate, completionDate and posting dates include their raw date, precision and ACTUAL/ESTIMATED type when supplied; missing dates or labels remain null.

The selected projection contains identity, title, sponsor, type, phases, conditions and status/date metadata. It excludes participant data, contacts, eligibility, protocol narratives, outcomes and linked documents. Source text is inert, with title capped at 500 characters, sponsor at 300, conditions/phases at 30 entries of 200 characters and status/type at 100 characters. Missing and truncation signals accompany the projection.

Attribution, processing date and freshness

Every operation fetches /version, selected study metadata, then /version again, without caching. Source API version and processing timestamp must match; a change fails with an explicit restart message and no automatic retry. dataTimestamp preserves the source UTC processing timestamp; dataProcessedAt explicitly adds the UTC designation. versionFetchedAt and fetchedAt identify separate observations. versionVerifiedAt identifies the second version observation. Matching versions reduce update races but are not an atomic study/search snapshot or cross-backend consistency guarantee; the processing date is not the study’s last update date. cacheHit is always false.

Attribute registry metadata to ClinicalTrials.gov and identify the selection and projection by ToolCargo. Keep the source processing date and modification disclosures with redistributed metadata, and refresh it when using it again rather than presenting an old copy as current. Current terms permit free worldwide access but retain international and third-party copyrights. ToolCargo claims no proprietary database ownership, blanket CC0 license, reuse clearance or government endorsement.

Provider and research limits

Requests use the fixed HTTPS ClinicalTrials.gov v2 metadata origin, an identifying support contact, shared provider-only lease, at least 1.1 seconds between upstream starts and source-directed cooldown. Each response is bounded to 2 MiB and 15 seconds; the operation is bounded to 35 seconds with 60-second abandoned-lease recovery. No automatic retries, redirect following, arbitrary links, assets, source accounts or ToolCargo credential forwarding. No numeric upstream quota or availability guarantee is claimed.

Study records are supplied by sponsors and investigators; limited registry review is not a determination of study safety or scientific merit. This connector supports public record research and does not provide treatment recommendations, eligibility assessment or regulatory, medical or legal guarantees.

Primary sources: API documentation · Current v2 OpenAPI specification · Terms and conditions · Disclaimer.