Clinicaltrials

What Clinicaltrials MCP server does

Clinicaltrials MCP server connects AI agents to the NIH-operated ClinicalTrials.gov public registry through its API v2. The registry contains clinical studies across phases 1–4, conditions, interventions, sponsors, locations, and post-market safety surveillance. The integration is independent and unofficial; it is not published or endorsed by ClinicalTrials.gov.

The core search flow uses ct_search to find studies by keyword, condition, drug or therapy, status, phase, or a named trial. Results include NCT identifiers, titles, recruitment status, enrollment, and sponsor information. ct_get_study retrieves a complete record for a known NCT ID, including protocol details, eligibility criteria, outcomes, locations, and posted results.

How it works

The server translates MCP tool calls into ClinicalTrials.gov API queries and returns structured study data. NCT IDs provide stable references for individual records, and the documented canonical URI format is pipeworx://clinicaltrials/study/{nct_id}.

Several tools support analysis across registry records. ct_count_by_condition groups trial counts by status and phase. ct_sponsor_trials lists studies for a registered lead sponsor, while ct_compare_sponsors compares two to five named sponsors under matching filters. ct_sponsor_pipeline summarizes a sponsor’s portfolio, and ct_competitive_landscape maps registered interventional studies across sponsors, phases, statuses, and interventions.

Time- and location-based tools cover different research workflows. ct_recent_updates distinguishes study edits, first registrations, and first-posted results within date ranges. ct_trials_by_location finds studies near a named place or coordinates, defaulting to recruiting status. ct_catalyst_calendar finds sponsor-entered completion dates. ct_sponsor_activity tracks a sponsor’s registrations, result postings, or updates.

Tools and capabilities

• ct_enrollment_watch identifies active trials for enrollment follow-up using registry fields and site counts.
• ct_results_summary summarizes registered endpoints, outcome measures, analyses, and adverse-event groups for one study.
• ct_phase_transition_rates estimates registry-derived Phase 1-to-2 and Phase 2-to-3 transitions and compares them with a BIO/Informa/QLS benchmark.
• Sponsor filters normally use the registered lead-sponsor field; an optional collaborator-inclusive mode is available where supported.
• Counts and comparisons can be constrained by condition, status, phase, intervention, sponsor, date, or location.

Limitations and notes

Clinicaltrials MCP server reports registry data, not independent medical or investment conclusions. Sponsor counts describe registered lead sponsorship and do not prove drug ownership, pipeline value, scientific differentiation, or likelihood of success. Registry completion dates may be estimated or revised, so they are planning signals rather than confirmed readouts, presentations, or regulatory events.

The phase-transition tool counts trial existence and observed registry progression, not efficacy, approval, or technical success. It reconstructs interventions from free-text names and applies censoring rules; combination regimens and common background therapies can affect the results. It returns insufficient-data status instead of a rate when fewer than ten eligible interventions are available.

A completed study does not necessarily have posted results, and many studies that publish in journals may not post results in the registry. Sponsor-name variations can change counts, while phase 2/3 studies may match both phases.