Make healthcare intelligence usable.

The U.S. healthcare market is not short on files, dashboards, or expert opinions. It is short on source-grounded intelligence that students, operators, researchers, builders, and communities can actually use.

Explore the portfolio View GitHub
Healthcare admin agents Public-data tools with provenance Profiles and open publishing
Built for trust and use

The intelligence stack.

A practical framework for turning public records into judgment, evidence, reference work, and competitive analysis.

Featured project

Healthcare Agents turns operator questions into structured workups.

The bench is written as long-form role files, scored against a frozen evaluation rubric, and designed to route administrative questions to the right specialist with explicit evidence needs and risk flags.

51specialists across the bench
10healthcare operating divisions
94.18average tracked eval
Healthcare Admin Workbench View repository
Alpha Healthcare Data MCP

The connector layer for the scattered public healthcare record.

Public healthcare data is everywhere and nowhere: CMS files, price transparency disclosures, research registries, sanctions lists, provider identifiers, grant databases, contract records, quality programs, demographic context, and market geography.

Healthcare Data MCP is a FastMCP toolkit that turns those sources into callable tools with structured responses, source metadata, bounded caches, and deployment modes for local agents or shared local processes.

Python FastMCP Local agents Shared process Structured responses Source metadata Bounded cache
  • 18 MCP servers and 100 tools in the current alpha release.
  • Designed for agent workflows that need sourced retrieval instead of hand-waved healthcare claims.
  • Honest limitation: the public data layer is incomplete, messy, and frequently delayed. The product does not pretend otherwise; it makes provenance visible.
18servers
100tools
v0.1.1release
6GitHub stars
2forks
0open issues
View repository
DATA CONNECTOR COVERAGE Public sources normalized for agent calls.

Tool responses carry source metadata, bounded cache behavior, and structured outputs instead of unsupported healthcare claims.

CMS and facility dataNPPES, quality, safety, star ratings.
Price and claims contextRates, volumes, service lines, case mix.
Public recordsSAM.gov, CHPL, 340B, breaches, exclusions.
Research and trialsNIH RePORTER, ClinicalTrials.gov profiles.
Market geographyDrive times, service areas, Census context.
Referral networksPhysician flows, leakage, network mapping.
REPOSITORY SIGNAL Alpha codebase, recently active.
Building United States Health Systems Observatory

The largest encyclopedia of comparative U.S. health system data.

USHSO is the synthesis layer: a living reference library that compares health systems as organizations, competitors, employers, capital allocators, care delivery networks, and public institutions.

Each report is designed to combine audited filings, CMS quality data, price transparency signals, service-line volumes, market geography, referral behavior, leadership changes, M&A activity, strategy, and public commitments into a single defensible profile.

System profiles Audited filings CMS quality Market geography Referral behavior Leadership changes Student access
  • The goal is breadth and depth: system-by-system intelligence that students, operators, researchers, founders, journalists, and communities can actually use.
  • Early coverage is intentionally narrow while the report format and data pipeline harden; Jefferson Health and Temple Health are the first visible examples.
  • Student access is free. The commercial model exists to fund maintenance without hiding the methodology.
2visible examples
Freestudent access
Sourcemethodology-first
Paidsupport model
Review evidence posture
PROFILE DOSSIER Comparative system intelligence with visible methodology.

Reports are meant to read like living dossiers: financial evidence, care quality, market position, and strategic movement in one place.

Financial performanceMargins, cash, debt, payer mix, capital allocation.
Quality and outcomesHCAHPS, readmissions, safety, star ratings.
Market positionShare, service areas, volumes, competitors.
Strategic directionLeadership, M&A, partnerships, expansion.
EARLY COVERAGE
Jefferson Health Temple Health Report format hardening Data pipeline hardening
Live American Journal of Healthcare Strategy

Open-access healthcare strategy needs serious publishing operations.

AJHCS is the public distribution partner in the portfolio: a journal, media platform, and knowledge institution built around open access to healthcare strategy, leadership, and management thought.

Open-Informatics supports organizations like AJHCS with website publishing, ranking, bookstore infrastructure, discoverability, and digital operations for free. That matters because open knowledge fails when the publishing layer looks amateur, fragile, or invisible.

Open access Podcasting Bookstore Reports Search Newsletter Institutional reach
  • AJHCS gives the portfolio a public venue for evidence-based essays, interviews, reports, podcasts, and practical healthcare strategy writing.
  • The infrastructure work is deliberately pro bono: the point is to make credible healthcare knowledge easier to publish and easier to find.
  • Honest limitation: publishing reach is earned over time. The site is a compounding asset, not a finished institution.
13K+subscribers
500Kmonthly impressions
200+podcast episodes
Visit AJHCS
PUBLISHING OPERATIONS Open healthcare strategy moves through a durable distribution stack.

AJHCS is treated as live infrastructure: publishing, packaging, and audience channels are supported as one operating surface.

AJHCS.org homepage preview showing open-access healthcare strategy publishing
Live AJHCS.org publishing surface.
INFRASTRUCTURE SIGNAL AJHCS.org is live and reachable. HTTP 200 verified; public preview shows the journal, podcast, bookstore, and report infrastructure in one place.
PublishEssays, interviews, reports, podcasts.
PackageBookstore, reports, feeds, archives.
DistributeSearch, social, newsletter, institutions.

Impressive is not the same thing as inflated.

The portfolio is ambitious because the problem is large. The claims stay concrete: working repositories, visible publishing infrastructure, specific data sources, defined agent domains, and an explicit admission that public healthcare data is imperfect.

Source-first.

Agents should be pushed toward public evidence, citations, and data retrieval instead of confident unsupported summaries.

Open where possible.

The tools and publishing infrastructure are designed to expand access, not create another closed advisory moat.

Competitive by design.

The end state is a market where more people can compare health systems, challenge assumptions, and build better healthcare organizations.