Clinical Catalyst Radar for Biotech

Find the biotech name where an asset is advancing through the clinic and management is buying the stock at the same time before the rest of the market connects the two.

For each company you pass, this pipeline joins three independent signals:

1. Clinical progress — ClinicalTrials.gov studies sponsored by the company: trial phase, recruiting/active status, enrollment, recent updates, posted results.
2. Research momentum — PubMed papers behind the program and their citation weight.
3. Insider conviction — SEC Form 4 open-market purchases (code P) by company insiders.

A trial advancing is public knowledge that can take quarters to reprice. Insider buying is informed but tells you nothing about why. The edge is the overlap: a Phase 2/3 asset moving and insiders putting their own cash in. This pipeline scores that convergence 0–100 and tiers each company.

Input

{
  "companies": ["Moderna | MRNA", "CRISPR Therapeutics | CRSP", "BioNTech | BNTX"],
  "insiderLookbackDays": 90,
  "clinicalSinceDays": 540,
  "trialsPerCompany": 25,
  "papersPerCompany": 20,
  "includeInsider": true,
  "minScore": 0
}

Each company is Company Name | TICKER. The name drives the ClinicalTrials.gov sponsor match and the PubMed search; the ticker drives the Form 4 insider lookup. Trials are attributed to a company only when its name matches the study's lead sponsor or a collaborator (corporate suffixes like Inc., Therapeutics, Pharma are ignored), so the clinical signal is the company's own pipeline, not a keyword soup.

Output

One row per company:

{
  "ticker": "MRNA",
  "company": "Moderna",
  "tier": "catalyst",
  "convergenceScore": 71,
  "scoreBreakdown": { "clinical": 34, "research": 22, "insider": 15 },
  "clinical": {
    "trialCount": 9,
    "topPhase": "Phase 3",
    "advancingCount": 4,
    "recentlyUpdatedCount": 5,
    "withResultsCount": 2,
    "topTrials": [{ "nctId": "NCT0...", "phase": "Phase 3", "status": "ACTIVE_NOT_RECRUITING", "sponsor": "ModernaTX, Inc.", "enrollment": 1200, "lastUpdate": "2026-05-12", "url": "https://clinicaltrials.gov/study/NCT0..." }]
  },
  "research": {
    "paperCount": 14,
    "recentPaperCount": 9,
    "totalCitations": 640,
    "highImpactCount": 3,
    "topPapers": [{ "pmid": "...", "year": 2025, "citations": 210, "title": "...", "url": "https://pubmed.ncbi.nlm.nih.gov/..." }]
  },
  "insider": {
    "buyCount": 3, "sellCount": 0, "buyValueUsd": 480000, "netValueUsd": 480000, "netBuying": true,
    "lastBuyDate": "2026-05-28",
    "topBuys": [{ "insider": "Jane Doe", "role": "Director", "shares": 5000, "price": 32.1, "valueUsd": 160500, "date": "2026-05-28" }]
  },
  "scoredAt": "2026-06-10T..."
}

Tiers

• catalyst — a Phase 2+ asset is actively advancing and insiders are net buyers, with a convergence score ≥ 55. The thesis is fully formed. The first catalyst per run is free.
• clinical_lead — a strong, advancing pipeline with research weight, but insiders are quiet.
• insider_lead — insiders are buying, but the clinical pipeline is early or thin.
• watch — limited signal on both sides.

Scoring (0–100)

• Clinical (0–45): highest trial phase (Phase 3/4 weighted near-market), count of actively advancing trials, recent updates, posted results.
• Research (0–30): recent (last 2 years) publication volume, total citations, count of above-expected-impact papers.
• Insider (0–25): open-market buy value, number of buys, net buying vs selling.

Pricing and cost

Pay per company, by tier: watch $0.05, single-signal lead $0.10, catalyst $0.16. The first catalyst per run is free so you can validate output.

This pipeline calls two of our own actors as children, and each child also bills you for its own per-row usage: pubmed-clinical-trials-intelligence (one charge per trial/paper row, called once per company) and sec-form4-insider-tracker (one charge per insider transaction, called once per run). Both children run on public government JSON APIs (NCBI E-utilities, ClinicalTrials.gov v2, SEC EDGAR) with no browser and no residential proxy, so the underlying compute is small and the per-company price stays above it.

Notes

• No API keys. Everything runs on public government endpoints.
• Attribution is per company. The clinical/research child is called once per company so trials and papers are matched to the right name, never guessed from a shared query.
• Graceful degradation. If one company's clinical lookup or the insider stage hits a transient block, that signal is skipped and the rest of the run still scores and returns.