# Pipeline analysis Machine path: GET /pipeline?sponsor= on this site, GET https://faers.mobi/signals for FAERS signals, and GET https://vaers.globalpatientsafety.com/signals for VAERS signals. GET /pipeline?sponsor= returns ClinicalTrials.gov studies for one company. format=brief is text/plain. format=html is the same page a browser shows. An unknown sponsor is an empty list, not an error. If the registry is down the response is 502 or 503 and contains no NCT ids. # Global Patient Safety API Machine path: GET /pipeline?sponsor= Registry facts from ClinicalTrials.gov. Not efficacy or safety conclusions. FAERS signals are a different service: https://faers.mobi/api VAERS signals are a different service: https://vaers.globalpatientsafety.com/api ## GET /pipeline sponsor (required for JSON and brief): trimmed, then matched against the lead sponsor name. A family name that contains it counts, such as Sanofi Pasteur for Sanofi. It is not expanded to other companies. role (optional): omitted means lead. collaborator means collaborator only. any means lead or collaborator, which is the ClinicalTrials.gov sponsor search. Each study has role lead or collaborator. n_studies counts that same role. status (optional): omitted means ongoing studies (RECRUITING, ENROLLING_BY_INVITATION, NOT_YET_RECRUITING, ACTIVE_NOT_RECRUITING). ALL means every status. Or one of RECRUITING, ENROLLING_BY_INVITATION, NOT_YET_RECRUITING, ACTIVE_NOT_RECRUITING, COMPLETED, SUSPENDED, TERMINATED, WITHDRAWN, UNKNOWN, AVAILABLE, NO_LONGER_AVAILABLE, TEMPORARILY_NOT_AVAILABLE, APPROVED_FOR_MARKETING, WITHHELD. phase (optional): EARLY_PHASE1, PHASE1, PHASE2, PHASE3, PHASE4, NA. limit (optional): default 50, maximum 200. pageToken (optional): the next_page_token from a previous response. format (optional): json, brief, or html. With format omitted, a client that sends Accept: text/html gets HTML. Any other client gets JSON. n_studies is the registry total. by_phase and by_status count only the studies in this response (counts_of is "returned"). An unknown sponsor is HTTP 200 with n_studies 0 and studies []. If ClinicalTrials.gov does not respond, the status is 503. If it returns an error, the status is 502. Those error bodies have error and source, and no NCT ids. ## Examples curl -sS 'https://globalpatientsafety.com/pipeline?sponsor=Sanofi' curl -sS 'https://globalpatientsafety.com/pipeline?sponsor=Sanofi&role=any' curl -sS 'https://globalpatientsafety.com/pipeline?sponsor=Sanofi&format=brief' curl -sS -H 'Accept: text/html' 'https://globalpatientsafety.com/pipeline?sponsor=Sanofi' OpenAPI: https://globalpatientsafety.com/openapi.json MCP: https://faers.mobi/mcp (tool get_pipeline). See /mcp.md Full page: https://globalpatientsafety.com/pipeline Signals are statistical patterns, not evidence of causation. Not medical advice. A pipeline report is a ClinicalTrials.gov registry fact, not a safety conclusion.