Constat

Constat is an FDA AI and machine learning medical-device database and evidence search platform available as a live console and through an MCP server. It is designed to help regulatory, quality, clinical-evidence, and market-access teams search source-linked records across the entire device lifecycle. The platform provides tools to analyze premarket clearances, track adverse events and recalls, monitor compliance actions, and trace reimbursement pathways. Users can also connect the evidence tools directly into external AI agent workflows.

• Search source-linked 510(k) summaries, validation designs, and predicate chains across 1,491 parsed devices
• Monitor MAUDE adverse event trends, recalls, re-clearances, and disclosed drift signals across 1,499 snapshotted devices
• Trace clearance pathways to payment codes, NTAP, CMS rates, and payer policies across 40 payment pathways
• Perform Radar category checks, firm history tracking, and monitored watchlist changes for FDA compliance
• Connect 13 source-linked FDA and CMS evidence tools inside ChatGPT, Claude, Codex, Cursor, VS Code, or custom MCP-compatible agents
• Track warning letters normalized by issue date, issuing center, firm, and violation subject with preserved source links
• Analyze device recalls containing FDA root-cause determinations for design, software, process, and suppliers

Pros

• Connects premarket 510(k) evidence directly to postmarket signals and reimbursement pathways within a single shared device corpus
• Provides machine-readable access via Model Context Protocol (MCP) to integrate public-record evidence directly into external developer environments and AI workflows
• Includes source-quoted fields and preserves links to primary FDA and CMS records for traceability
• Offers structured team services including scoped evidence assessments and configured monitoring briefs tied to specific regulatory decisions

Cons

• Anonymous MCP reads and public console searches are rate-limited on the free tier
• The platform provides descriptive decision-support based on statistical pattern analysis rather than formal legal, regulatory, coding, or reimbursement advice
• Higher-volume MCP access and licensed evidence exports require contacting the vendor for a custom quote rather than straightforward self-serve signup
• Monitored evidence briefs and evidence assessments require scoping discussions before delivery begins

Constat fits regulatory affairs, quality, postmarket surveillance, clinical-evidence, market-access, and data/agent engineering teams needing to research FDA AI/ML medical devices, compare predicate chains, and track compliance or reimbursement data. It is a poor fit for individuals or organizations looking for guaranteed legal, regulatory, coding, or reimbursement advice rather than descriptive public-record intelligence.

Constat offers a free tier providing public console search and anonymous, rate-limited Model Context Protocol (MCP) access with no credit card required. Paid options include the Evidence Assessment starting at a one-time fee from $5,000 for a bounded, source-linked review around a single decision delivered within a 2-week target. A Monitored Evidence Brief is available starting from $500 per month to maintain a configured watchlist for agreed product codes, firms, competitors, or suppliers. Evidence Data and MCP access for higher-volume usage and licensed data exports use usage-based pricing that requires contacting the vendor.

Unlike generic search engines or manual FDA database lookups, Constat specifically aggregates AI and ML medical-device data into a single shared corpus linking premarket 510(k) records, postmarket signals, and reimbursement paths. It bridges traditional regulatory research with modern developer workflows by exposing 13 source-linked evidence tools directly through Model Context Protocol (MCP) servers. Rather than simply returning raw database rows, it pairs automated ingestion with structured outputs that keep attribution, source links, and interpretation limits visible.