You want clean, cross-linked data — not another scraping project. ClinicaLister pre-links trials, drugs, devices, substances, and US/EU registries with the match method and a confidence score attached, ready to export for your analysts or read straight from an AI agent.
The highest-leverage things ClinicaLister does for a data manager / bioinformatician.
Every recurring activity, mapped to what ClinicaLister does — on which surface (the App, the MCP graph, or both), the time it saves, and how confident that claim is.
| Activity | What ClinicaLister does | Surface | Time saved | Confidence |
|---|---|---|---|---|
| Build CT.gov↔FDA↔EMA↔GSRS↔GUDID↔NDC pipeline | The cross-source join you'd hand-build arrives pre-computed — trials, drugs, devices, substances, and registries already linked across 18+ sources. | Both | weeks → hours; core pain removed | High |
| Entity resolution to UNII / CAS / formula | Every substance arrives with its UNII, molecular formula, CAS number, and weight, reconciled across US and EU records. | Both | no hand-built crosswalk | High |
| Device entity resolution | Device identifiers link to their trials automatically, replacing the manual UDI matching. | Both | manual GUDID matching removed | High |
| US↔EU cross-registry reconciliation | US and EU trial and drug records come pre-matched, skipping a notoriously messy join. | Both | skips a notoriously messy join | Medium |
| Data QC / trusting a join | Every link shows how it was matched and how confident that match is, so QC is a filter-and-spot-check, not a re-derivation. | Both | QC = filter + spot-check | High |
| Provenance / audit trail | The matching method and confidence travel with each link, so audit and reproducibility metadata comes built in. | Both | audit metadata comes free | High |
| Cohort / dataset assembly for scientists | A relationship-rich slice — trials, molecules, diseases, and their links — comes out as one cross-linked export. | Both | relationship-rich slice in one export | High |
| Molecule enrichment (targets / class / variants) | Your AI assistant can walk the molecule graph by target, drug class, or variant without a hand-built lookup. | MCP | ready-made graph traversal | Medium |
| Feed analytics / BI warehouse | One cross-linked export feeds your BI tools and pipelines as periodic loads. | App | periodic loads — not warehouse-native | Medium |
| Deliver dataset to analysts | One click produces a cross-linked workbook with links and custom columns included — an immediate hand-off artifact. | App | immediate hand-off artifact | High |
| Wire data into AI / agent workflows | Your AI assistant and agents read the data natively over MCP — a first-class agent and RAG path, no export layer to build. | MCP | native agent / RAG path | High |
| Scheduled refresh + delta reconciliation | The data refreshes through the day, and you can pull just what's new or changed since your last run. | Both | delta tooling — pull, not push | Medium |
| Large programmatic / bulk extraction | Not supported — there's no public REST or bulk-data API to mirror the full dataset. | — | — (blocked, see gaps) | n/a |
Straight answers on where ClinicaLister stops today — so there are no surprises.
No public REST or bulk-data API — the one load-bearing constraint. You can't mirror the whole dataset into a lake, wire up a warehouse/ELT connector, stream changes, or run arbitrary high-throughput pulls; the data is meant to be exported or read by an agent, not bulk-replicated.
Links are confidence-scored suggestions, not ground truth. The final validation before anything becomes authoritative is still yours.
Not a live feed. The data is pull-based — you poll for what's new; there's no push, webhook, or change-stream for near-real-time reactive pipelines.
Export-shaped, not query-shaped. There's no SQL access, so any join beyond the cross-linked sheets provided is reconstructed on your side after export.
Coverage follows the public record. Undisclosed programs and non-public identifiers won't appear.
Trials, drugs, devices, substances, and US/EU registries arrive already linked — each connection tagged with how it was matched and how confident that match is. Ready to export for your analysts or query straight from an AI agent. Stop building pipelines; start analyzing.