FDA Intelligence MCP Server

Search FDA regulatory data across 9 openFDA endpoints using 11 MCP tools -- covering drug approvals, adverse events, recalls, medical device clearances, PMA approvals, MAUDE reports, food safety events, and more. No API key needed. Over 46 million records accessible.

This MCP server is built for regulatory analysts, pharma researchers, medical device evaluators, food safety teams, and anyone who needs programmatic access to FDA data through AI assistants like Claude, Cursor, Windsurf, or any MCP-compatible client.

Quick Start

Claude Desktop

Add this to your Claude Desktop configuration file (claude_desktop_config.json):

{
  "mcpServers": {
    "fda-intelligence": {
      "url": "https://actors-mcp-server.apify.com/sse#ryanclinton/fda-intelligence-mcp"
    }
  }
}

Cursor

In Cursor Settings, go to MCP and add a new server with this URL:

https://actors-mcp-server.apify.com/sse#ryanclinton/fda-intelligence-mcp

Windsurf / Other MCP Clients

Use the SSE connection URL:

https://actors-mcp-server.apify.com/sse#ryanclinton/fda-intelligence-mcp

Once connected, your AI assistant will have access to all 11 FDA intelligence tools. Try asking:

• "What is the adverse event profile for Ozempic?"
• "Show me Class I drug recalls from Pfizer"
• "Find 510(k) clearances for glucose monitors"
• "Run a full regulatory intelligence report on metformin"

Available Tools

#	Tool	Description	Key Parameters
1	fda_search_drug_approvals	Search NDA/ANDA/BLA drug approvals	drug_name, sponsor, active_ingredient, limit
2	fda_drug_adverse_events	Get drug adverse event aggregate statistics from FAERS (27M+ reports)	drug_name, manufacturer, date_from, date_to, top_n
3	fda_search_drug_recalls	Search drug recall enforcement reports	drug_name, recalling_firm, classification, reason_keyword, status, date_from, date_to, limit
4	fda_search_device_clearances	Search 510(k) premarket device clearances (230K+ records)	device_name, applicant, product_code, date_from, date_to, limit
5	fda_search_device_approvals	Search PMA Class III device approvals (56K+ records)	trade_name, applicant, advisory_committee, date_from, date_to, limit
6	fda_device_adverse_events	Get MAUDE device adverse event statistics (19M+ reports)	device_name, manufacturer, date_from, date_to, top_n
7	fda_search_device_recalls	Search device recall enforcement reports	device_name, recalling_firm, classification, date_from, date_to, limit
8	fda_food_adverse_events	Get CAERS food/supplement adverse event statistics (148K+ reports)	product_name, date_from, date_to, top_n
9	fda_search_food_recalls	Search food recall enforcement reports	product_name, recalling_firm, classification, reason_keyword, date_from, date_to, limit
10	fda_product_intelligence	Full cross-domain regulatory intelligence report with risk assessment	product_name, company_name, scope, limit_per_category
11	fda_list_sources	List all available endpoints and data coverage	(none)

Tool Details

1. fda_search_drug_approvals

Search the Drugs@FDA database for NDA, ANDA, and BLA drug approvals by drug name, sponsor, or active ingredient.

When to use: Finding when a drug was approved, who manufactures it, what formulations exist.

Example prompt: "Search for FDA approvals of metformin"

Example response:

{
  "total": 34,
  "results": [
    {
      "applicationNumber": "NDA021842",
      "sponsorName": "TAKEDA PHARMS USA",
      "brandNames": ["ACTOPLUS MET"],
      "genericNames": ["PIOGLITAZONE HYDROCHLORIDE AND METFORMIN HYDROCHLORIDE"],
      "activeIngredients": ["PIOGLITAZONE HYDROCHLORIDE", "METFORMIN HYDROCHLORIDE"],
      "dosageForms": ["TABLET"],
      "routes": ["ORAL"],
      "productCount": 4,
      "submissionCount": 22,
      "originalApprovalDate": "2009-10-21"
    }
  ]
}

2. fda_drug_adverse_events

Get aggregate safety statistics from the FDA Adverse Event Reporting System (FAERS) -- total reports, serious/death counts, and top adverse reactions. Uses count endpoints to summarize millions of reports without fetching individual records.

When to use: Understanding a drug's safety profile, comparing adverse reaction patterns, tracking safety signals.

Example prompt: "What are the adverse event statistics for Ozempic?"

Example response:

{
  "drugName": "Ozempic",
  "totalReports": 95230,
  "seriousReports": 72145,
  "deathAssociatedReports": 1842,
  "seriousnessRate": "75.8%",
  "topAdverseReactions": [
    { "term": "NAUSEA", "count": 18432 },
    { "term": "DIARRHOEA", "count": 9812 },
    { "term": "VOMITING", "count": 8754 },
    { "term": "ABDOMINAL PAIN", "count": 6291 },
    { "term": "CONSTIPATION", "count": 4832 }
  ],
  "dateRange": { "from": "all", "to": "present" },
  "note": "Adverse event reports do not prove causation. High counts may reflect widespread usage."
}

3. fda_search_drug_recalls

Search drug recall enforcement reports. Filter by drug name, recalling firm, classification (Class I/II/III), reason keyword, or status.

When to use: Finding recall history for a drug or manufacturer, investigating contamination events.

Example prompt: "Find Class I drug recalls involving NDMA contamination"

Example response:

{
  "total": 47,
  "results": [
    {
      "recallNumber": "D-0741-2020",
      "classification": "Class I",
      "status": "Terminated",
      "recallingFirm": "Aurobindo Pharma USA Inc",
      "productDescription": "Valsartan Tablets, 160 mg, Rx only",
      "reasonForRecall": "NDMA detected above acceptable daily intake limit",
      "recallInitiationDate": "2019-05-14",
      "distributionPattern": "Nationwide",
      "brandNames": ["VALSARTAN"]
    }
  ]
}

4. fda_search_device_clearances

Search the FDA 510(k) premarket clearance database for medical devices cleared via the 510(k) pathway (substantially equivalent to existing devices).

When to use: Finding cleared devices by name or manufacturer, checking regulatory status.

Example prompt: "Search for 510(k) clearances for glucose monitors"

Example response:

{
  "total": 342,
  "results": [
    {
      "kNumber": "K221345",
      "deviceName": "CONTINUOUS GLUCOSE MONITORING SYSTEM",
      "applicant": "DEXCOM INC",
      "decisionDate": "2023-03-15",
      "decision": "SUBSTANTIALLY EQUIVALENT (SE)",
      "productCode": "MDS",
      "reviewAdvisoryCommittee": "Clinical Chemistry",
      "statementOrSummary": "Summary"
    }
  ]
}

5. fda_search_device_approvals

Search the PMA (Premarket Approval) database for Class III high-risk medical device approvals. These devices require full clinical evidence for approval.

When to use: Finding PMA approvals for implants, life-sustaining devices, or other high-risk Class III devices.

Example prompt: "Find PMA approvals for cardiovascular devices from Medtronic"

Example response:

{
  "total": 892,
  "results": [
    {
      "pmaNumber": "P860023",
      "tradeName": "IMPLANTABLE CARDIOVERTER DEFIBRILLATOR",
      "genericName": "IMPLANTABLE DEFIBRILLATOR",
      "applicant": "MEDTRONIC INC",
      "decisionDate": "2023-08-22",
      "decisionCode": "APPR",
      "advisoryCommittee": "CV",
      "supplementNumber": "S425"
    }
  ]
}

6. fda_device_adverse_events

Get aggregate device adverse event statistics from the MAUDE (Manufacturer and User Facility Device Experience) database.

When to use: Understanding device safety signals, comparing adverse event patterns across device types.

Example prompt: "Show me adverse event statistics for hip implants"

Example response:

{
  "totalReports": 45230,
  "topDeviceCategories": [
    { "term": "HIP PROSTHESIS", "count": 28412 },
    { "term": "TOTAL HIP REPLACEMENT", "count": 12341 }
  ],
  "reportTypes": [
    { "term": "Malfunction", "count": 22150 },
    { "term": "Injury", "count": 18432 },
    { "term": "Death", "count": 312 }
  ],
  "query": { "device_name": "hip implant", "manufacturer": null }
}

7. fda_search_device_recalls

Search medical device recall enforcement reports by device name, firm, classification, or date range.

When to use: Finding device recall history, evaluating manufacturer recall patterns.

Example prompt: "Show me device recalls from Abbott Laboratories"

8. fda_food_adverse_events

Get aggregate food and dietary supplement adverse event statistics from the CAERS (Center for Food Safety and Applied Nutrition Adverse Event Reporting System) database. Covers food, dietary supplements, cosmetics, and infant formula.

When to use: Investigating supplement safety, tracking food-related adverse events, cosmetic reactions.

Example prompt: "What adverse events are reported for green tea extract supplements?"

Example response:

{
  "productName": "green tea extract",
  "totalReports": 312,
  "topReactions": [
    { "term": "LIVER INJURY", "count": 87 },
    { "term": "ABDOMINAL PAIN", "count": 45 },
    { "term": "NAUSEA", "count": 38 }
  ],
  "topOutcomes": [
    { "term": "HOSPITALIZATION", "count": 124 },
    { "term": "VISITED AN ER", "count": 89 }
  ],
  "dateRange": { "from": "all", "to": "present" }
}

9. fda_search_food_recalls

Search food recall enforcement reports. Filter by product, firm, classification, or reason keyword.

When to use: Finding food recall history, investigating Salmonella, E. coli, Listeria, or undeclared allergen events.

Example prompt: "Find food recalls due to Salmonella contamination in 2024"

10. fda_product_intelligence

Generate a comprehensive cross-domain FDA regulatory intelligence report. Queries all relevant endpoints in parallel (up to 9) and returns a unified risk-assessed profile with cross-domain timeline.

When to use: Full regulatory due diligence, M&A target assessment, product safety deep dive.

Example prompt: "Run a full FDA intelligence report on Medtronic"

Example response (abbreviated):

{
  "query": { "product_name": null, "company_name": "Medtronic", "scope": "all" },
  "summary": {
    "riskLevel": "HIGH",
    "domainsWithData": ["drugs", "devices"],
    "totalApprovals": 47,
    "totalAdverseEvents": 312450,
    "totalRecalls": 23,
    "classIRecalls": 2,
    "deathReports": 1842
  },
  "drugs": { "approvals": [...], "adverseEvents": {...}, "recalls": [...] },
  "devices": {
    "clearances510k": [...],
    "approvalsPma": [...],
    "adverseEvents": { "totalReports": 289000, "topDeviceProblems": [...] },
    "recalls": [...]
  },
  "food": null,
  "timeline": [
    { "date": "2025-11-10", "domain": "device", "type": "clearance", "description": "510(k) K241234 -- CARDIAC PACEMAKER" },
    { "date": "2025-09-15", "domain": "device", "type": "recall", "description": "Class I recall -- Software defect in pacemaker firmware" }
  ]
}

Risk levels computed automatically:

Level	Criteria
CRITICAL	Multiple Class I recalls AND >100 death reports, OR Class I recalls AND >50 deaths
HIGH	Any Class I recall OR >50 death-associated reports
MODERATE	Recalls present AND >1,000 adverse events, OR any recalls with >100 adverse events
LOW	Some recalls or adverse events present but not reaching higher thresholds
MINIMAL	Few or no concerning findings across all domains

11. fda_list_sources

List all available FDA data sources, endpoints, record counts, and coverage details. No parameters required.

Use Cases

• Pharma due diligence -- Before acquiring a drug portfolio, get the full regulatory profile: approvals, recall history, adverse event trends, and computed risk level
• Medical device evaluation -- Research a device's 510(k) clearances, PMA approvals, MAUDE adverse event reports, and recall history before procurement or investment
• Drug safety research -- Investigate a drug's FAERS adverse event profile including top reactions, death counts, and seriousness rates
• Food safety monitoring -- Track food recalls by pathogen (Salmonella, Listeria, E. coli) or investigate supplement adverse events in CAERS
• NDMA/contamination investigation -- Trace recalls and adverse events for drugs affected by impurity contaminations (valsartan, ranitidine, metformin)
• Competitor intelligence -- Search by company name to map a competitor's full FDA regulatory footprint across drugs, devices, and food
• Regulatory compliance monitoring -- Integrate with scheduled workflows to track evolving risk profiles for products in your portfolio
• Legal research -- Build a complete regulatory timeline for litigation involving drug or device safety claims
• Supply chain risk assessment -- Evaluate manufacturer recall history and adverse event patterns before sourcing decisions

Data Coverage

Endpoint	Tool(s)	Records	Description
drug/drugsfda	fda_search_drug_approvals	~29K	NDA, ANDA, and BLA drug approval applications
drug/event	fda_drug_adverse_events	~27M	FAERS drug adverse event reports (aggregate statistics)
drug/enforcement	fda_search_drug_recalls	~17K	Drug recall enforcement reports
device/510k	fda_search_device_clearances	~230K	510(k) premarket device clearances
device/pma	fda_search_device_approvals	~56K	PMA Class III device approvals
device/event	fda_device_adverse_events	~19M	MAUDE device adverse event reports (aggregate statistics)
device/enforcement	fda_search_device_recalls	~38K	Device recall enforcement reports
food/event	fda_food_adverse_events	~148K	CAERS food/supplement adverse event reports (aggregate statistics)
food/enforcement	fda_search_food_recalls	~25K	Food recall enforcement reports

Total: 46M+ records accessible across 9 openFDA endpoints.

How It Works

  AI Assistant (Claude, Cursor, etc.)
           │
           │  MCP Protocol (SSE)
           ▼
  ┌────────────────────────────────────┐
  │  FDA Intelligence MCP Server       │
  │  (Apify Actor in Standby Mode)     │
  │                                    │
  │  11 registered MCP tools           │
  │  Express + StreamableHTTP          │
  └──────────────┬─────────────────────┘
                 │
    ┌────────────┼────────────┐
    │            │            │
    ▼            ▼            ▼
  DRUGS       DEVICES       FOOD
  ───────     ────────      ────
  drugsfda    510k          food/event
  drug/event  pma           food/enforcement
  drug/enf    device/event
              device/enf
    │            │            │
    └────────────┼────────────┘
                 │
                 ▼
         openFDA API
        (api.fda.gov)
         Free & Public

1. Your AI assistant connects to the MCP server via the Apify MCP gateway
2. The server exposes 11 tools that the assistant can call based on your natural language queries
3. Each tool constructs optimized openFDA API queries with proper field searches and count aggregations
4. For adverse event tools, the server uses openFDA's count endpoints to aggregate millions of records into top-N summaries without fetching individual reports
5. The fda_product_intelligence tool queries up to 9 endpoints in parallel and computes a risk assessment
6. Automatic retry with exponential backoff handles rate limits (429 responses)
7. Results are returned as structured JSON through MCP back to your assistant

Pricing

This actor uses Apify's pay-per-event billing via Actor.charge(). Each tool call is a separate billing event. The openFDA API itself is completely free -- you only pay for the Apify actor compute.

Tool	Event	Approximate Cost
fda_search_drug_approvals	drug-approvals	~$0.005
fda_drug_adverse_events	drug-adverse-events	~$0.005
fda_search_drug_recalls	drug-recalls	~$0.005
fda_search_device_clearances	device-clearances	~$0.005
fda_search_device_approvals	device-approvals	~$0.005
fda_device_adverse_events	device-adverse-events	~$0.005
fda_search_device_recalls	device-recalls	~$0.005
fda_food_adverse_events	food-adverse-events	~$0.005
fda_search_food_recalls	food-recalls	~$0.005
fda_product_intelligence	product-intelligence	~$0.01
fda_list_sources	(free)	$0

A typical research session of 5-10 tool calls costs well under $0.10.

Limitations

• No API key required -- but openFDA enforces rate limits (240 requests per minute per IP, 120K per day). The server handles 429 responses with automatic retry and exponential backoff.
• Result caps -- openFDA limits search results to 1,000 per query. For products with more records, results are a representative sample. Aggregate statistics (adverse event counts, top reactions) cover the full dataset.
• Adverse event reports are not proof of causation -- a drug listed in a FAERS report may not have caused the reaction. High death counts often reflect widespread usage (e.g., metformin has 34K+ death reports because it is prescribed to millions of patients).
• Not all records have openFDA enrichment -- some enforcement reports lack brand_name or generic_name cross-references, which can affect search completeness.
• Food adverse event searches require a product name -- the food/event endpoint does not support company-name-only queries.
• Date filters apply to different fields per endpoint -- receivedate for drug events, decision_date for device clearances, report_date for enforcement, etc.
• Data freshness -- openFDA data is typically updated weekly by the FDA. The server queries live data on every call.

Related Actors

This MCP server combines the same data sources available through these individual actors in the FDA regulatory intelligence suite:

Actor	What It Does
FDA Drug Approval Search	Detailed drug approval records with full field access
openFDA Drug Event Monitor	Individual FAERS adverse event reports with full detail
FDA Drug Recall Search	Drug recall enforcement reports with full filtering and pagination
FDA 510(k) Device Clearances	Medical device 510(k) clearance records
FDA PMA Device Approvals	Class III PMA device approval records
FDA Device Adverse Events (MAUDE)	Medical device adverse event reports from MAUDE
FDA Medical Device Recall Search	Device recall enforcement reports
FDA Food & Supplement Adverse Events	CAERS food/supplement adverse event reports
FDA Food Recall Monitor	Food recall enforcement reports
FDA Product Intelligence	Standalone actor version of the cross-domain intelligence report

Use the individual actors when you need raw data extraction with full field access and pagination. Use this MCP server when you want your AI assistant to interactively query FDA data through natural language conversation.