The sources behind the watch.

How we use it. Discovers development-stage assets from a ClinicalTrials.gov mirror and records them as pipeline molecules, with their research codes and registry trials.

Why it matters. The only source that finds assets before they are marketed — the FDA NDC catalog can only contain products that already have an NDC.

How we use it. We use ClinicalTrials.gov to track U.S. and international clinical studies for each medicine, including status, phase, sponsor, enrollment, key dates, and posted results.

Why it matters. It helps users understand development progress, upcoming catalysts, and the evidence base behind a product.

How we use it. We use public Chronos gene effects to summarize whether cancer cell lines depend on a drug target.

Why it matters. It adds a functional screen alongside genetics, tumor alterations, and target-disease associations.

Context. Cell-line results are preclinical and should not be described as patient response evidence.

How we use it. We use the EU Clinical Trials Information System to capture studies registered for European review.

Why it matters. It adds regional visibility that can reveal development activity not obvious from U.S. registries alone.

How we use it. We use epidemiology sources to estimate disease burden, prevalence, incidence, mortality, and patient populations where source coverage allows.

Why it matters. It supports addressable-population estimates, burden-of-disease context, and market-sizing work.

Context. Some sources have usage restrictions; Watchtower only enables them when appropriate rights are configured.

How we use it. We use FDA Federal Register notices to track regulatory-review periods and patent-term-restoration applications for specific products.

Why it matters. It adds lifecycle-extension context while keeping applicant-requested days distinct from the term USPTO ultimately awards.

Context. Requested extension days are not the term ultimately awarded by USPTO. Notices are product/application-specific.

How we use it. Reconciles known patent publications to DOCDB-compatible publication numbers, simple-family IDs, global family members, abstracts, dates, assignees, inventors, and CPC/IPC classifications.

Why it matters. Replaces title/applicant family guessing with source-provided family identity and adds global publication context.

Context. CC BY 4.0. Queries are credential-gated, project-billed, and capped by GOOGLE_PATENTS_MAXIMUM_BYTES_BILLED. Family expansion starts from an already linked publication or family.

How we use it. We use published genetic associations and population constraint to add human-biology context for a medicine's targets.

Why it matters. It helps users distinguish target hypotheses with human-genetic context from those supported only by experimental or mechanistic evidence.

Context. A mapped GWAS gene is not automatically causal, and genetic constraint is not proof of treatment benefit or safety.

How we use it. We use NIH grant records to identify publicly funded U.S. projects, principal investigators, institutions, institutes, and award years.

Why it matters. It shows where U.S. public research momentum exists and which experts or institutions are active around a therapy area.

Context. Authoritative U.S./NIH grants source; OpenAlex Awards adds non-NIH funders separately.

How we use it. We use Open Targets to connect medicines with diseases, targets, known drugs, and biological evidence.

Why it matters. It helps users evaluate target rationale, nearby mechanisms, safety considerations, and related development programs.

Context. Some biology and indication evidence is molecule-level, so it may apply across related brands.

How we use it. We use OpenAlex Awards to capture research grants from funders beyond NIH, including international public agencies, foundations, and industry programs.

Why it matters. It broadens the funding picture to non-U.S. and private funders so research momentum is not limited to NIH-funded work.

Context. Award amounts come in many currencies and are not directly comparable across funders.

How we use it. Normalizes OpenAlex work topics, citation/reference edges, influence metrics, and optional author profile stats/topics from existing publication and person records.

Why it matters. Adds queryable research themes, citation lineage, top-percentile publication signals, and KOL-level OpenAlex influence context.

Context. Work enrichment is local from stored OpenAlex payloads; author profile hydration calls the OpenAlex Authors API only when enabled.

How we use it. Retrieves article full text, preferring NLM's PMC datasets and falling back to crawling open publisher URLs; stores text with its section structure and license.

Why it matters. Supports per-arm results, subgroup and safety detail, methods, and funding/disclosure evidence that abstracts cannot answer.

Context. Mixes the PMC Open Access Subset with the PMC Author Manuscript dataset. Author manuscripts are licensed for text and data mining only and are recorded as such, so downstream surfaces do not reproduce them at length.

How we use it. Re-reads retrieved full text to extract funding, disclosure, and contribution evidence and to re-run drug mention extraction.

Why it matters. Turns retrieved articles into queryable provenance and mention evidence, and re-runs anything a dropped background job missed.

How we use it. Streams NLM's daily DOI/PMCID/PMID crosswalk and resolves the PMC accession for publications we already hold.

Why it matters. Unlocks the fast full-text path: a publication without a PMC accession cannot be retrieved from the PMC article datasets.

How we use it. We use ROR to normalize institutions, funders, hospitals, companies, and research organizations found across publications, grants, trials, and other sources.

Why it matters. It turns messy organization strings into stable identities with aliases, geography, organization type, and hierarchy context.

Context. ROR identifies organizations; it does not by itself imply product sponsorship, funding, or endorsement.

How we use it. Builds direct and composed research lineage edges across grants, publications, trials, and patents from existing source-backed links and identifiers.

Why it matters. Supports translation-path questions such as grant-to-publication lag, foundational publications, trial-publication linkage, and publication-to-patent signals.

Context. Derived edges preserve provenance and confidence; related-work and shared-person patent links are lower-confidence leads, not proof of causality.

How we use it. We use publication indexes to connect medicines with papers, preprints, authors, institutions, topics, and abstracts.

Why it matters. It supports evidence review, key opinion leader discovery, and fast orientation around the science behind an asset.

How we use it. Matches exact molecule InChIKeys to patent publications whose CLAIMS name the compound, then imports section-aware compound occurrences, DOCDB simple-family IDs, assignees, publication dates, and classifications from the biweekly Parquet snapshot.

Why it matters. Adds deterministic molecule-level patent discovery and distinguishes compounds found in claims from compounds found only in descriptions, abstracts, titles, images, or MOL attachments.

Context. CC BY 4.0. A compound occurrence is extracted disclosure evidence, not proof of claim scope or exclusivity. Assignees reflect publishing-authority data and may not include later reassignments. Coverage requires an exact InChIKey on the molecule, so peptides and other biologics are generally out of scope. Linking is claims-gated, and structures in more than 5,000 claims patents are skipped as ubiquitous chemistry.

How we use it. We use PTAB proceeding and decision records to show post-grant patent challenges, petitioners, status, and published Board decisions.

Why it matters. It adds a direct litigation-risk signal that patent grants and Orange Book dates alone cannot provide.

Context. These records are molecule-level patent events, not legal conclusions about a medicine's full patent estate.

How we use it. We use USPTO application and certificate workbooks to track filed restoration requests and extensions actually awarded to patents.

Why it matters. It distinguishes a requested extension from the USPTO certificate that determines the awarded period.

Context. These are molecule-level patent records; year-based awards are shown exactly as USPTO publishes them.

How we use it. We use patent records to surface applications, patent numbers, inventors, applicants, dates, and patent status related to a medicine.

Why it matters. It helps users understand intellectual-property activity, lifecycle strategy, and potential exclusivity signals.

How we use it. We use verified chemical identifiers to recognize the same active ingredient across databases and related brands.

Why it matters. It makes cross-source matching more reliable and reduces duplicate or mistaken molecule identities.

Context. These identifiers belong to the active-ingredient molecule, not exclusively to one brand.

How we use it. We use the WHO trial registry search to broaden coverage across national and regional trial registries.

Why it matters. It helps users see a more global clinical-development picture and reduces blind spots outside the U.S. and EU.

How we use it. We use public TCGA cohort data to show mutations, amplifications, and deep deletions in drug target genes.

Why it matters. It reveals tumor contexts where a target alteration is observed and preserves the denominator behind each frequency.

Context. Study-cohort frequency is not general population prevalence or proof that the tumor depends on the target.

How we use it. We use ClinVar to connect clinically relevant genetic variants with targets when that evidence is available.

Why it matters. It adds biomarker and precision-medicine context for targeted therapies.

How we use it. We use DDInter to identify known drug-drug interaction pairs and severity context.

Why it matters. It helps users spot interaction considerations that may matter for prescribing, trial design, or competitive positioning.

How we use it. We use DailyMed labels to capture current and historical prescribing information, including safety, dosing, indications, and label changes.

Why it matters. It grounds product views in official labeling and makes label evolution easier to review.

How we use it. We use EMA medicine data to track European authorizations, holders, approval dates, and medicine classifications.

Why it matters. It helps users compare U.S. and European regulatory status and identify regional launch or approval differences.

How we use it. We use FDA's quarterly AEMS/FAERS reports to track potential serious safety signals and FDA's subsequent evaluation or action.

Why it matters. It distinguishes FDA-reviewed surveillance signals from raw spontaneous adverse-event report volume.

Context. A listed potential signal does not prove the medicine caused the event. Ingredient-level signals can apply across sibling brands.

How we use it. We use FDA accelerated-approval listings to track conditional approvals, confirmatory-study expectations, conversions, and withdrawals.

Why it matters. It helps users understand where approval is tied to future evidence and where withdrawal risk may exist.

How we use it. We use FDA adverse-event data to summarize reported reactions and recent safety reports connected to a medicine.

Why it matters. It provides safety-signal context for risk review and follow-up investigation.

Context. Adverse-event reports are not incidence rates and do not prove causation.

How we use it. We use FDA advisory committee calendars and materials to identify public review meetings tied to medicines or therapeutic areas.

Why it matters. It gives users early visibility into regulatory events, briefing materials, and public discussion points.

How we use it. We use FDA approval and Orange Book records to track applications, submissions, approval history, marketing status, exclusivity, and product/application patent listings.

Why it matters. It anchors regulatory timelines and helps users understand how a product reached market and what source-listed protections remain.

Context. Orange Book patent expiration dates are sponsor-submitted FDA listing data, not legal opinions; practical entry timing can change.

How we use it. We use FDA companion-diagnostic records to connect therapies with required or cleared tests and biomarkers.

Why it matters. It clarifies testing dependencies that can shape eligible patients, adoption, and launch execution.

How we use it. We use FDA complete response letter data to identify public regulatory setbacks and the FDA concerns connected to them.

Why it matters. It helps users understand approval risk, resubmission context, and issues that can shape future regulatory strategy.

How we use it. We load FDA label text across the branded catalog to record what each product is officially indicated for, plus dosing, contraindications, and warnings.

Why it matters. It separates what a medicine is approved to treat from what it has merely been studied in, which is the difference between a label and a trial record.

How we use it. We use FDA shortage records to track availability issues, reasons, presentations, and timing.

Why it matters. It helps users understand supply constraints that may affect patients, providers, and competitors.

Context. Shortage records can apply at the molecule or presentation level, not always to one brand alone.

How we use it. We use FDA snapshots to show who participated in original-approval trials and FDA's subgroup benefit and safety summaries.

Why it matters. It adds product-specific demographic representation and subgroup context.

Context. Snapshots apply only to original approval and are not updated for later uses; the older program index is loaded from the archive linked by FDA.

How we use it. We use FDA Paragraph IV, approval, authorized-generic, and no-approved-generic records to identify generic-entry pressure and market structure.

Why it matters. It helps users distinguish a patent challenge, an approval, an authorized-generic entry, and an application with no approved generic.

How we use it. We use FDA package and product identifiers to connect medicines with public pricing, utilization, coverage, and package-level records.

Why it matters. It improves matching accuracy across public-program datasets and helps avoid mixing brands that share the same generic ingredient.

Context. Brand pages stay brand-specific; shared molecule data is handled separately when needed.

How we use it. We use FDA orphan records to track designation, approval, withdrawal or revocation, sponsor, and orphan exclusivity at the indication level.

Why it matters. It adds rare-disease development and commercial-protection context beyond a product's current approval record.

Context. An orphan designation is not an FDA marketing approval. Ingredient-level rows can apply across sibling brands.

How we use it. We use FDA postmarketing records to track studies and commitments required or agreed to after approval.

Why it matters. It shows what evidence a sponsor still needs to generate after launch and whether obligations are completed or pending.

How we use it. We use FDA product-specific guidance records to show published generic-development methods and planned new or revised guidance.

Why it matters. It helps users assess generic-development readiness, complexity, and upcoming methodology changes.

Context. Planned dates can change and do not mean a guidance is already available.

How we use it. We use Purple Book records to track biologic licenses, reference products, biosimilars, and interchangeability status.

Why it matters. It helps users understand biologic competition and follow-on product dynamics.

How we use it. We use FDA REMS records to identify products with risk evaluation and mitigation programs.

Why it matters. It highlights safety-management requirements that may affect prescribing, distribution, patient access, and commercialization.

How we use it. We use FDA recall and enforcement records to track product-quality events, recall reasons, firms, classifications, and dates.

Why it matters. It surfaces supply-chain and quality signals that can affect product perception or availability.

How we use it. We use FDA's SrLC database to show approved safety-label actions, affected sections, supplement dates, and amended language.

Why it matters. It distinguishes an authoritative FDA safety-label action from a generic change detected between DailyMed label versions.

Context. FDA SrLC coverage begins in January 2016.

How we use it. We discover official brand, patient-support, savings, HCP, and prescribing-information pages for each medicine.

Why it matters. It gives users quick access to the public product materials that often explain access programs, support resources, and prescribing details.

How we use it. We use RxNorm to standardize medicine names, identifiers, and drug-class context.

Why it matters. It makes search, matching, and cross-source comparisons more consistent.

How we use it. Uses the Plan Attributes PUF as the medical-plan inventory, then queries the CMS Marketplace API by product-level RxCUIs that RxNorm resolved from trusted brand NDCs.

Why it matters. Shows ACA Marketplace plan listing by issuer, state, market, metal level, and plan type with explicit Covered, NotCovered, GenericCovered, and DataNotProvided states.

Context. The API does not publish tier, prior authorization, step therapy, or quantity limits. This is not employer coverage, generally excludes state exchanges that do not use the federal platform, and requires CMS_MARKETPLACE_API_KEY.

How we use it. We use public CMS metadata to surface proposed LCDs, national analyses, MEDCAC meetings, and technology assessments that explicitly name a medicine or molecule.

Why it matters. It provides earlier visibility into evidence review and possible future Medicare policy activity.

Context. A linked development record is not a coverage decision and must be read in its source context.

How we use it. We use public CMS policy metadata to surface NCDs, LCDs, and articles that explicitly name a Part B medicine or molecule.

Why it matters. It adds policy timing and national-versus-local context without reproducing licensed CPT descriptions.

Context. A linked policy still needs to be read for indication, setting, and jurisdiction-specific coverage.

How we use it. We use annual and preliminary quarterly Part B spending data to track fee-for-service spending, claims, beneficiaries, and dosage units by HCPCS code.

Why it matters. It shows utilization and Medicare spend for physician-administered products, which ASP price files alone cannot provide.

Context. Some HCPCS codes cover multiple products; those rows are not presented as exclusive to one brand.

How we use it. We use the latest official CMS monthly release to track Part D plan listings, tiers, prior authorization, step therapy, quantity limits, and plan geography by exact product NDC.

Why it matters. It establishes source-specific Medicare Part D access and makes the difference between listed, not listed in a completed release, and not yet evaluated explicit.

Context. Medicaid PDL and broad commercial formulary coverage are not supplied by this source.

How we use it. We use Medicare Part D spending data to track public-program spend, claims, beneficiaries, and utilization trends.

Why it matters. It helps users understand commercial scale and payer exposure in Medicare.

How we use it. We use monthly plan enrollment to weight Part D formulary listings and restrictions by beneficiary population.

Why it matters. It distinguishes broad access in large plans from the same count of listings in very small plans.

Context. Suppressed enrollment of 10 or fewer is excluded rather than estimated.

How we use it. We use Open Payments to identify reported financial relationships between manufacturers, clinicians, teaching hospitals, and research organizations.

Why it matters. It helps users understand relationship networks, research funding, and potential KOL connections.

Context. Research payments can be molecule-level and may apply across related brands.

How we use it. We use annual ownership and investment-interest files to identify interests held by physicians or their immediate family members in reporting entities.

Why it matters. It adds conflict-of-interest context to physician and company intelligence.

Context. These records are company relationships, not drug-level payments. Cumulative value-of-interest snapshots are shown one program year at a time.

How we use it. We use provider/service rows for matched drug HCPCS codes to summarize NPI, specialty, geography, site of care, services, beneficiaries, and Medicare payment amounts.

Why it matters. It reveals provider concentration and site-of-care patterns for physician-administered products.

Context. Beneficiary counts below 11 are suppressed by CMS, and service counts are billing units rather than unique administrations.

How we use it. We use Part D prescriber data to summarize prescribing by provider, specialty, geography, claims, and cost.

Why it matters. It reveals where use is concentrated and which specialties are most involved.

How we use it. Extracts royalty tiers, milestone ladders, and upfronts from filed material contracts.

Why it matters. Gives deal comparables real contract economics instead of headline totals, flagging terms filed under confidential treatment.

Context. Model-based extraction; runs on demand or on an opt-in schedule.

How we use it. Fetches narrative exhibits attached to filings — earnings press releases (EX-99) and material contracts (EX-10).

Why it matters. Opens up the substance behind a filing's cover page, where guidance and real contract terms actually live.

How we use it. Extracts forward revenue, EPS, and margin guidance from earnings press releases.

Why it matters. Anchors near-term projections to management's own stated outlook rather than modeled assumptions.

Context. Model-based extraction; runs on demand or on an opt-in schedule.

How we use it. Extracts the gross-to-net revenue bridge issuers disclose in their revenue-reserve footnotes.

Why it matters. Turns gross published pricing into a net price, the single largest correction between list and realized revenue.

Context. Model-based extraction; runs on demand or on an opt-in schedule.

How we use it. Parses Section 16 ownership filings into insider purchases, sales, grants, and option exercises.

Why it matters. Shows management and large-holder conviction, separating discretionary trades from compensation mechanics.

How we use it. We use NADAC pricing to track public acquisition-cost benchmarks for pharmacy-dispensed medicines.

Why it matters. It adds pricing context for public-program reimbursement and market-access analysis.

How we use it. We use Medicaid utilization data to track state-level prescriptions, units, and reimbursement by quarter.

Why it matters. It helps users understand public-program utilization and geographic variation in Medicaid.

How we use it. We use Medicare negotiation data to identify selected drugs, negotiated prices, and effective years.

Why it matters. It flags products with direct policy exposure and helps users assess future pricing pressure.

How we use it. We use Medicare Part B pricing files for payment-limit benchmarks on physician-administered products.

Why it matters. It helps users understand reimbursed pricing context for products covered under Part B.

How we use it. We use the NHS Business Services Authority's monthly English Prescribing Dataset to track prescription items, dispensed quantity, and NHS cost for a molecule and its brands in England's primary care.

Why it matters. It is a timely, monthly demand signal that complements slower US public-program data.

Context. Contains public sector information licensed under the Open Government Licence v3.0. Prescribing is molecule-level; generic-written items cannot be assigned to one brand. Items are prescriptions, not patients.

How we use it. Ingests issuer-tagged us-gaap and ifrs-full financial facts — cash, operating cash flow, revenue, expenses, and balance sheet — in each SEC-reported currency.

Why it matters. Provides exact reported financials and a same-currency cash-runway read for funding and dilution risk.

How we use it. We use search interest trends to monitor public awareness and demand signals for medicine names.

Why it matters. It helps users spot changes in attention around launches, news events, safety issues, or competitive moments.

How we use it. We use broad news and press sources to track market, clinical, regulatory, and competitive coverage around medicines.

Why it matters. It surfaces timely signals that may appear before they are reflected in structured regulatory or commercial datasets.

How we use it. We use brand-name episode searches to discover podcast coverage and retain links back to the podcast and its publisher feed.

Why it matters. It adds long-form clinician, patient, and industry discussion that may not appear in news or structured datasets.

Context. A title or description match is metadata evidence, not proof that the brand was spoken. Spoken mentions require an available transcript.

How we use it. We use company filings to connect medicines with reported revenue, risks, corporate events, and management disclosures.

Why it matters. It grounds commercial and strategic context in issuer-filed source documents.

Context. The automatic universe combines SEC life-sciences SIC tiers with tracked/watched companies and drug full-text discoveries. Primary documents, full-text-matched attachments, and narrative exhibits retain separate identities; attachment bodies are fetched selectively.

How we use it. Discovers and refreshes the SEC filer universe across core therapeutics/biotech and adjacent diagnostics, device, medical-laboratory, and testing-laboratory SIC tiers.

Why it matters. Extends company diligence beyond the current drug catalog while preserving tracked, watched, and drug-full-text-discovered companies whose SEC industry classification is incomplete.

Context. Coverage records the number of CIKs discovered, profiles loaded, and profile failures by SIC; a failed EDGAR page marks the refresh failed rather than silently returning partial coverage.

How we use it. We use public social and video content, when available, to capture brand-oriented discussion and attention signals.

Why it matters. It adds patient, clinician, and public-interest context that structured datasets can miss.

Context. YouTube transcript matches are spoken-word and noisier than text; generic/molecule-term hits in transcripts are molecule-level and must not be summed across sibling brands.