Food Safety Supply Chain MCP Server

Food safety supply chain intelligence via the Model Context Protocol — giving AI assistants direct access to FDA recalls, adverse event reports, supplier hygiene ratings, ingredient trade flows, contamination pathways, and seasonal risk projections. Built for food manufacturers, restaurant chains, importers, and food safety consultants who need live regulatory and supply chain data in their AI workflows.

This MCP server orchestrates 7 public data sources in parallel using a single tool call. Each query fans out to the FDA recall database, FDA CAERS adverse event system, UK Food Hygiene registry, Open Food Facts, UN COMTRADE trade statistics, OpenAQ air quality monitors, and NOAA weather alerts — then synthesises the results through 4 dedicated scoring models into a structured risk report. No subscription, no manual searches: pay only per tool call.

What data can you extract?

Why use this food safety MCP?

Manual food safety monitoring means logging into the FDA recall portal, searching CFSAN's adverse event system, cross-referencing UK Food Standards Agency records, pulling UN COMTRADE reports, and then correlating all of that by hand. For a single ingredient or supplier assessment, this takes 3–5 hours and still leaves blind spots — you may catch the recall but miss the trade flow pattern that predicted it.

This MCP automates the entire process. A single tool call collects data from all 7 sources simultaneously and applies scoring models that weight Class I recalls at 10 points each, flag biological pathogens (Salmonella, Listeria, E. coli, Campylobacter, Norovirus, Botulism) at 8 points per detection, penalise poor hygiene ratings, and correlate recall frequency against import volume to surface systemic risk.

• Scheduling — run daily or weekly food safety sweeps for monitored ingredients to keep risk profiles current
• API access — trigger assessments from Python, JavaScript, or any HTTP client inside your quality management workflow
• Proxy rotation — all underlying data collection uses Apify's built-in proxy infrastructure for reliable regulatory database access
• Monitoring — get Slack or email alerts when supply chain risk reports return HIGH or CRITICAL ratings
• Integrations — connect to Zapier, Make, Google Sheets, or quality management systems via webhooks

Features

• 7 parallel data sources — FDA Food Recall Monitor, FDA CAERS adverse events, UK Food Hygiene, Open Food Facts, UN COMTRADE, OpenAQ, and NOAA Weather all queried simultaneously per tool call
• Ingredient Risk Heat Map — correlates UN COMTRADE import volume with FDA recall and adverse event frequency, calculating a recall-per-million-dollar-trade-volume correlation score (max 20 points)
• Biological pathogen detection — scans recall reasons for 8 named pathogens: Salmonella, Listeria, E. coli, Campylobacter, Norovirus, Botulism, Clostridium, and variants
• Chemical hazard detection — identifies 9 chemical contamination categories in recall text: allergens, undeclared ingredients, lead, mercury, arsenic, pesticide residues, melamine, aflatoxin, and sulfites
• Environmental contamination scoring — compares PM2.5 and PM10 readings from OpenAQ against WHO food safety thresholds (PM2.5 >15 µg/m³, PM10 >45 µg/m³) near production zones
• UK Food Hygiene composite — rates establishments on the 0–5 FSA scale, applies a volume-weighted score, penalises poor ratings (0–1 stars), and adds Open Food Facts Nutri-Score data as a product quality signal
• Seasonal risk projection — tracks monthly recall distribution to compute a current-month-versus-average ratio, amplified by NOAA heat alert severity (EXTREME = 10 pts, SEVERE = 6 pts)
• Four-model composite scoring — Ingredient Risk (30%) + Contamination Pathways (25%) + Supplier Hygiene inverted (25%) + Seasonal Risk (20%) = composite 0–100 score
• Four-level risk classification — LOW, MODERATE, HIGH, CRITICAL with configurable thresholds at 25/50/75
• Actionable recommendations — rule-based recommendations generated when specific score thresholds are crossed: HACCP review triggers at 3+ contamination pathways; supplier audit triggers at CRITICAL ingredient risk
• Standby mode — runs as a persistent HTTP server, meaning tool calls respond in seconds without cold start delays
• Pay-per-event pricing — $0.045 per tool call, no subscription, no minimum spend

Use cases for food safety supply chain intelligence

Food manufacturer ingredient sourcing

Quality assurance teams at food manufacturers need to evaluate supplier risk before contracting. Use trace_ingredient_risk to cross-reference every proposed ingredient with FDA recall history and UN COMTRADE import volume data. An ingredient arriving from a country with high recall correlation per trade dollar is a risk signal that warrants additional supplier audit — surfaced in seconds, not days.

Restaurant chain compliance monitoring

Multi-location restaurant groups and franchise operators need continuous hygiene compliance visibility. Use assess_supplier_hygiene with location or business names to pull UK FSA inspection scores across a supplier network. The Supplier Hygiene Composite Score flags establishments drifting below acceptable hygiene levels before a regulatory inspection forces action.

Food import and customs risk profiling

Importers and brokers managing high-volume food shipments can use trace_ingredient_risk with a country code to profile origin-country risk for specific commodity categories. Correlating UN COMTRADE volume with FDA recall density identifies which import corridors carry the highest per-unit recall risk, enabling targeted pre-import testing.

Allergen and adverse event signal detection

Regulatory affairs teams tracking emerging safety signals use analyze_adverse_events to monitor FDA CAERS reports for specific ingredients or brands. The outcome distribution analysis surfaces hospitalisations and life-threatening events before they aggregate into a formal recall, enabling proactive reformulation or label review.

Seasonal supply chain planning

Procurement and operations teams planning summer buying cycles use project_seasonal_risk to compare current NOAA heat alert severity against historical FDA recall seasonality patterns. When the current month shows 40%+ above-average recall frequency alongside extreme heat advisories, the model flags PEAK seasonal risk — time to increase cold chain monitoring frequency.

Contamination incident investigation support

Food safety investigators responding to consumer complaints use detect_contamination_pathways to triage three simultaneous contamination vectors. Biological pathogen hits in recent recall data, chemical hazard patterns in adverse event reports, and elevated PM2.5 near a production facility can all be assessed in a single call, narrowing investigation scope before lab results return.

How to run food safety supply chain queries

1. Connect the MCP server — add the server URL https://food-safety-supply-chain-mcp.apify.actor/mcp to your MCP client (Claude Desktop, Cursor, Windsurf, or any compatible client). Provide your Apify API token as the Bearer token in the Authorization header.
2. Choose your tool — start with generate_supply_chain_risk_report for a full assessment of any ingredient or supplier, or use focused tools like search_food_recalls or assess_supplier_hygiene for targeted checks.
3. Provide a query — type a food product name ("peanut butter"), ingredient ("shrimp"), business name ("Pinnacle Foods Manchester"), or category ("dairy supplements"). Optionally add a location for weather and air quality context.
4. Read the structured results — the server returns JSON with risk scores, contamination pathways, hygiene levels, actionable recommendations, and all supporting raw records from underlying data sources. Copy into a report, push to a spreadsheet, or feed downstream AI analysis.

Input parameters

This MCP server exposes its interface as MCP tools rather than actor input fields. There is no traditional input schema — all parameters are passed as tool call arguments.

Input tips

• Start with generate_supply_chain_risk_report — this single tool calls all 7 data sources and all 4 scoring models at once, giving the most complete risk picture for $0.045.
• Use focused tools for speed — if you only need recall data, search_food_recalls runs one actor instead of seven and returns faster.
• Add coordinates for environmental risk — supply latitude and longitude to detect_contamination_pathways for a precise 25 km radius air quality query around a specific production facility.
• Specify country for import risk — passing a UN country code (e.g., "156" for China, "704" for Vietnam) to trace_ingredient_risk narrows trade flow data to that origin, sharpening the recall correlation.
• Filter recalls by class — use classification: "Class I" in search_food_recalls to see only the most serious recalls (risk to health or life) without Class II/III noise.

Output example

Below is a representative output from generate_supply_chain_risk_report for the query "shrimp":

{
  "query": "shrimp",
  "compositeScore": 58,
  "riskLevel": "HIGH",
  "ingredientRisk": {
    "score": 72,
    "recallCount": 9,
    "adverseEventCount": 14,
    "tradeFlowVolume": 2400000000,
    "riskCorrelation": 3.75,
    "riskLevel": "HIGH",
    "topRecallReasons": [
      { "reason": "Salmonella contamination", "count": 4 },
      { "reason": "Undeclared allergens", "count": 3 },
      { "reason": "Listeria monocytogenes", "count": 2 }
    ],
    "signals": [
      "9 FDA food recalls — elevated ingredient risk",
      "4 serious adverse events (death/hospitalization)"
    ]
  },
  "contaminationPathways": {
    "score": 54,
    "biologicalRisk": 24,
    "chemicalRisk": 18,
    "environmentalRisk": 12,
    "pathwayCount": 5,
    "pathways": [
      "Biological: salmonella",
      "Biological: listeria",
      "Chemical: allergen",
      "Chemical: undeclared",
      "Environmental: air particulate contamination"
    ],
    "signals": [
      "Multiple biological pathogen detections in recall history",
      "Chemical contamination pathways detected (allergens/metals/pesticides)",
      "Environmental air quality exceeds food safety thresholds"
    ]
  },
  "supplierHygiene": {
    "score": 62,
    "establishmentsChecked": 38,
    "averageRating": 4.1,
    "fiveStarPct": 57.9,
    "zeroStarPct": 2.6,
    "hygieneLevel": "GOOD",
    "ratingDistribution": { "0": 1, "1": 2, "2": 3, "3": 5, "4": 8, "5": 19 },
    "signals": []
  },
  "seasonalRisk": {
    "score": 48,
    "activeWeatherAlerts": 6,
    "severeAlerts": 3,
    "currentSeasonRisk": "MODERATE",
    "weatherFactors": [
      "Extreme heat — cold chain disruption risk",
      "Flooding — water contamination risk"
    ],
    "signals": [
      "3 severe/extreme weather alerts — food safety impact"
    ]
  },
  "allSignals": [
    "9 FDA food recalls — elevated ingredient risk",
    "4 serious adverse events (death/hospitalization)",
    "Multiple biological pathogen detections in recall history",
    "Chemical contamination pathways detected (allergens/metals/pesticides)",
    "Environmental air quality exceeds food safety thresholds",
    "3 severe/extreme weather alerts — food safety impact"
  ],
  "recommendations": [
    "Multiple contamination pathways identified — review HACCP plan",
    "Biological pathogen risk elevated — enhance microbiological testing",
    "Peak seasonal risk — increase testing frequency and cold chain monitoring"
  ]
}

Output fields

How much does it cost to run food safety supply chain queries?

This MCP uses pay-per-event pricing — you pay $0.045 per tool call. Platform compute costs are included. There is no monthly fee, no minimum spend, and no data limits per call.

You can set a maximum spending limit per run to control costs. The actor stops when your budget is reached, protecting against runaway usage in automated workflows.

Compare this to food safety information services like FoodLogiQ or SafetyChain at $500–2,000/month — with this MCP, most teams running daily ingredient checks spend under $5/month with no subscription commitment. The Apify free tier includes $5 of monthly platform credits, covering over 100 tool calls at no cost.

How to connect this food safety MCP server

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "food-safety-supply-chain": {
      "url": "https://food-safety-supply-chain-mcp.apify.actor/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_TOKEN"
      }
    }
  }
}

Cursor / Windsurf / Cline

Add the server to your MCP settings under the Tools or MCP section:

• URL: https://food-safety-supply-chain-mcp.apify.actor/mcp
• Auth header: Authorization: Bearer YOUR_APIFY_TOKEN

Python (via HTTP)

import requests

response = requests.post(
    "https://food-safety-supply-chain-mcp.apify.actor/mcp",
    headers={
        "Content-Type": "application/json",
        "Authorization": "Bearer YOUR_APIFY_TOKEN"
    },
    json={
        "jsonrpc": "2.0",
        "method": "tools/call",
        "params": {
            "name": "generate_supply_chain_risk_report",
            "arguments": {
                "query": "peanut butter",
                "location": "California"
            }
        },
        "id": 1
    }
)
report = response.json()["result"]["content"][0]["text"]
import json
data = json.loads(report)
print(f"Composite risk score: {data['compositeScore']} ({data['riskLevel']})")
for rec in data.get("recommendations", []):
    print(f"  - {rec}")

JavaScript

const response = await fetch("https://food-safety-supply-chain-mcp.apify.actor/mcp", {
    method: "POST",
    headers: {
        "Content-Type": "application/json",
        "Authorization": "Bearer YOUR_APIFY_TOKEN"
    },
    body: JSON.stringify({
        jsonrpc: "2.0",
        method: "tools/call",
        params: {
            name: "trace_ingredient_risk",
            arguments: {
                ingredient: "shrimp",
                country: "704"
            }
        },
        id: 1
    })
});
const data = await response.json();
const result = JSON.parse(data.result.content[0].text);
console.log(`Ingredient risk: ${result.ingredientRisk.score} — ${result.ingredientRisk.riskLevel}`);
console.log(`Trade correlation: ${result.ingredientRisk.riskCorrelation} recalls per $M`);

cURL

# Full supply chain risk report
curl -X POST "https://food-safety-supply-chain-mcp.apify.actor/mcp" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_APIFY_TOKEN" \
  -d '{"jsonrpc":"2.0","method":"tools/call","params":{"name":"generate_supply_chain_risk_report","arguments":{"query":"spinach","location":"California","country":"484"}},"id":1}'

# Search recalls only
curl -X POST "https://food-safety-supply-chain-mcp.apify.actor/mcp" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_APIFY_TOKEN" \
  -d '{"jsonrpc":"2.0","method":"tools/call","params":{"name":"search_food_recalls","arguments":{"query":"romaine lettuce","classification":"Class I"}},"id":2}'

How Food Safety Supply Chain MCP works

Phase 1: Parallel data collection

When a tool is called, the server dispatches concurrent requests to up to 7 Apify actors using Promise.allSettled. Each actor runs in a separate 256 MB container with a 120-second timeout. Individual actor failures are handled gracefully — a failed actor returns an empty array rather than blocking the entire response, so partial results are always returned. The actors queried are: fda-food-recall-monitor (FDA recall database), fda-food-adverse-events (FDA CAERS), uk-food-hygiene (FSA registry), open-food-facts (product database), un-comtrade-search (trade flows), openaq-air-quality (PM2.5/PM10 readings), and noaa-weather-alerts (NOAA alert feed).

Phase 2: Scoring model application

Four independent scoring functions process the collected data:

Ingredient Risk applies a weighted formula: recall severity (Class I = 10 pts, Class II = 5 pts, Class III = 2 pts, max 35) + adverse event burden with serious-outcome multiplier (max 25) + log-scaled trade volume (max 20) + recall-per-million-dollar correlation (max 20).

Contamination Pathways scans recall reason text for 8 biological pathogens and 9 chemical hazard keywords, awarding 8 points per biological hit (max 30) and 6 points per chemical hit (max 30), then adds WHO-threshold air quality exceedances (PM2.5 >15 µg/m³ = 5 pts, PM10 >45 µg/m³ = 3 pts, max 20) and adverse event severity amplification (max 20).

Supplier Hygiene uses the FSA 0–5 rating scale: average rating × 10 (max 50) + five-star percentage bonus (max 20) − poor rating penalty for 0–1 star establishments (max −30) + log2-scaled volume reliability bonus (max 15) + Open Food Facts Nutri-Score quality adjustment (grade A = +5, grade D/E = −2, max 15).

Seasonal Risk computes a monthly recall ratio comparing current-month recall count against the rolling average across all observed months, then adds NOAA weather severity scores (EXTREME = 10, SEVERE = 6, MODERATE = 3) and heat alert amplification (max 20 for sustained heat events).

Phase 3: Composite scoring and recommendations

The composite score weights the four models: Ingredient Risk (30%) + Contamination Pathways (25%) + Hygiene inverted (25%, since high hygiene = low risk) + Seasonal Risk (20%). Five recommendation rules are evaluated against threshold conditions: CRITICAL ingredient risk triggers an immediate supplier audit recommendation; 3+ contamination pathways triggers a HACCP plan review; POOR or BELOW_AVERAGE hygiene triggers alternative supplier consideration; PEAK seasonal risk triggers increased testing frequency; biological risk ≥20 triggers enhanced microbiological testing.

Phase 4: MCP response

All structured results are serialised to JSON and returned as MCP tool content. The server runs in Apify's Standby mode, keeping the Express HTTP server alive between requests to eliminate cold start latency. Each tool call is charged as a pay-per-event billing event before actor calls are dispatched, with budget enforcement at the event level.

Tips for best results

1. Use generate_supply_chain_risk_report for initial assessments. It queries all 7 data sources in one call for $0.045. Use focused tools only when you need to drill into a specific dimension without re-running the full pipeline.

2. Add geographic context to sharpen environmental scores. The contamination pathway model produces richer environmental risk signals when you provide latitude and longitude rather than a text location — this enables the OpenAQ actor to query a precise 25 km radius around the facility rather than a city name.

3. Filter search_food_recalls by Class I for triage. Class I recalls represent a "reasonable probability" of serious health consequences. Filtering to Class I alone removes noise and surfaces the highest-priority events for daily monitoring workflows.

4. Interpret the Supplier Hygiene Score directionally. The score is higher for better hygiene — unlike other scores where higher means more risk. A score of 80+ means EXCELLENT; below 40 means AVERAGE or worse. Check hygieneLevel alongside the numeric score.

5. Combine with Company Deep Research for supplier due diligence. The hygiene composite tells you about FSA inspection history; deep research adds financial health, litigation, and operational background on the same supplier entity.

6. Schedule daily runs for monitored ingredients. Use Apify's built-in scheduling to call trace_ingredient_risk each morning for your top 10 ingredients. At $0.045 per call, a daily 10-ingredient sweep costs $0.45/day — less than a fraction of the cost of any commercial food safety SaaS.

7. Use currentSeasonRisk: "PEAK" as a trigger. Wire the seasonal risk output to a webhook that alerts your quality team when currentSeasonRisk moves to PEAK. This creates an automated early-warning system for cold chain and pathogen risk season without manual monitoring.

Combine with other Apify actors

Limitations

• FDA data covers US-regulated products only. Recalls and adverse events from the EU, Canada, Australia, or other jurisdictions are not included. For EU-regulated products, UK Food Hygiene data provides a partial signal but does not cover the full EFSA recall system.
• UK Food Hygiene covers England, Wales, and Northern Ireland. Scotland operates a separate Food Standards Scotland system and is not included. Non-UK food establishments have no hygiene rating data available through this MCP.
• Trade flow data (UN COMTRADE) has a reporting lag. COMTRADE statistics are typically 6–18 months behind the current date. This means very recent trade shifts are not reflected in ingredient risk scores. Use as a directional signal, not a real-time tracking tool.
• OpenAQ coverage varies by geography. Air quality monitoring stations are unevenly distributed globally. Facilities in rural areas or developing countries may return no air quality data, leaving environmental contamination scores at zero rather than reflecting actual conditions.
• Contamination pathway detection relies on recall reason text. If a recall reason is vague (e.g., "potential adulteration" without specifying the contaminant), specific pathogen or chemical flags will not be triggered even if an actual hazard exists.
• Scoring models do not replace HACCP or food safety audits. Scores indicate risk signals from public data; they do not constitute an official safety assessment, regulatory certification, or legal opinion on product safety.
• NOAA weather data is US-focused. The weather alert feed covers US territories. Supply chains with production in Asia, South America, or Africa will not get weather disruption signals for those regions.
• No authentication or access restrictions on public data. All data sources are public government and open databases. Private supplier audit records, proprietary testing results, and internal inspection notes are not accessible.

Integrations

• Zapier — trigger food safety risk reports when new supplier records are added to your database and route HIGH/CRITICAL results to quality management channels
• Make — build automated workflows that run weekly ingredient risk traces and populate Google Sheets dashboards with current scores
• Google Sheets — push supply chain risk reports for all monitored ingredients into a live spreadsheet for food safety teams
• Apify API — call any MCP tool directly from quality management systems, ERP platforms, or custom food safety applications
• Webhooks — get immediate notifications when riskLevel returns CRITICAL, enabling rapid response before a recall affects your supply chain
• LangChain / LlamaIndex — include this MCP server in RAG pipelines to give AI agents live food safety data for supply chain analysis and report generation

Troubleshooting

• Low or zero contamination pathway scores despite known product issues — The detection engine matches against recall reason text. If the FDA recall database contains no records for the queried product or ingredient, scores will be low. Try broader search terms (e.g., "dairy" instead of a specific brand) or search for the manufacturer name directly in search_food_recalls.

• Supplier Hygiene Score near zero despite reputable businesses — UK Food Hygiene data is geographically limited to England, Wales, and Northern Ireland. If you are assessing suppliers outside this region, the uk-food-hygiene actor returns no rated establishments and the hygiene score reflects only Open Food Facts product data. The establishmentsChecked field will show 0 when this is the case.

• Trade flow data returning empty for an ingredient — UN COMTRADE uses standardised commodity codes (HS codes). Generic ingredient names may not match the COMTRADE query format directly. Try using the commodity category name (e.g., "crustaceans" instead of "shrimp") or the HS chapter description.

• Seasonal risk showing LOW despite summer heat — The seasonal model scores heat only when NOAA has active EXTREME or SEVERE heat alerts. If no active alerts exist at query time, the heat amplification score is zero even if temperatures are high. Check NOAA's current alert status for your region separately if conditions suggest elevated risk.

• Tool call returning a spending limit error — Your Apify account has reached the per-run spending limit. Increase the limit in your run settings or upgrade your Apify plan. The free tier includes $5 of monthly credits, covering over 100 tool calls.

Responsible use

• This MCP server only accesses publicly available government databases and open food product data.
• All FDA, UK FSA, UN COMTRADE, OpenAQ, and NOAA data is released under open government or public domain licences.
• Risk scores are analytical outputs derived from public data and do not constitute regulatory determinations, legal opinions, or official food safety certifications.
• Do not use this tool as a substitute for required regulatory inspections, HACCP plans, or certified food safety audits.
• For guidance on web scraping legality, see Apify's guide.

FAQ

How many food safety data sources does this MCP query in a single call? The generate_supply_chain_risk_report tool queries all 7 data sources simultaneously in a single $0.045 call: FDA Food Recall Monitor, FDA CAERS adverse events, UK Food Hygiene, Open Food Facts, UN COMTRADE, OpenAQ, and NOAA Weather. Focused tools query 1–3 sources depending on the tool.

How accurate is the food safety risk score for a given ingredient? The composite score is derived from public regulatory data and correlates ingredient recall history with import trade volume. It performs well for widely-traded commodities with sufficient recall history. Niche ingredients with few or no FDA records may return low scores not because they are low risk, but because regulatory data is sparse. Always treat scores as risk signals to investigate, not definitive safety ratings.

Does this MCP detect food recalls before they are officially announced? No. The MCP queries the FDA's published recall database and CAERS adverse event reports, which are official post-announcement records. However, adverse event signals — particularly accumulating hospitalisations and life-threatening outcomes — can surface emerging safety issues before they reach formal recall status if the pattern is detectable in CAERS data.

Is food recall data from this MCP legally usable for compliance documentation? The data is sourced from the FDA's official public database and UK FSA public records. However, using this MCP does not constitute formal regulatory compliance or certification. Consult your food safety legal counsel before citing MCP outputs in regulatory filings or compliance documentation.

How is this different from the FDA's own recall database website? The FDA website provides search and browsing but no API, no scoring, no cross-referencing with trade flow data, and no combination with adverse events or hygiene ratings. This MCP provides a structured JSON API, four scoring models, parallel multi-source correlation, and integration with 6 additional data sources — all accessible programmatically from any MCP-compatible AI client.

Can I search food safety data for products sold in Europe? Partially. UK Food Hygiene data covers England, Wales, and Northern Ireland. Open Food Facts includes products from EU markets. However, the EFSA (European Food Safety Authority) recall system and EU RASFF alerts are not currently included. FDA data is US-market-only.

How often is the recall and adverse event data updated? Data is fetched live at query time directly from the FDA's databases. There is no caching. Each tool call reflects the current published state of the FDA recall database at the moment of the call.

Can I use this MCP to monitor a list of ingredients automatically? Yes. Use Apify's scheduling feature to run trace_ingredient_risk or generate_supply_chain_risk_report on each ingredient in your list on a daily or weekly schedule. Combine with webhooks to alert your team when any ingredient's risk level moves to HIGH or CRITICAL.

What food safety pathogens does the contamination pathway model detect? The biological detection engine scans recall reason text for 8 pathogens: Salmonella, Listeria, E. coli (and E.coli variants), Campylobacter, Norovirus, Botulism, and Clostridium. The chemical detection engine covers 9 hazard categories: allergens, undeclared ingredients, lead, mercury, arsenic, pesticide residues, melamine, aflatoxin, and sulfites.

Is it legal to use public FDA and UK FSA data this way? Yes. FDA data is published as open government data under US federal government works policy (public domain). UK Food Standards Agency data is released under the Open Government Licence. UN COMTRADE data is public UN statistical data. All data sources used by this MCP are publicly accessible and legally available for programmatic use. See Apify's guide on web scraping legality.

How is this different from food safety SaaS tools like FoodLogiQ or SafetyChain? Commercial platforms offer workflow management, supplier portal features, audit templates, and team collaboration tools alongside data — and cost $500–2,000/month. This MCP provides the live regulatory data intelligence layer only, accessible from any AI assistant for $0.045 per query. It is best used to augment existing workflows with live FDA and regulatory data rather than as a replacement for a full quality management platform.

Can I combine this MCP with other MCP servers in Claude Desktop? Yes. You can add multiple MCP servers to your claude_desktop_config.json. Combining this server with a general web search MCP or a company research MCP allows your AI assistant to cross-reference food safety risk data with news, supplier background, and market context in a single conversation.

Help us improve

If you encounter issues, you can help us debug faster by enabling run sharing in your Apify account:

1. Go to Account Settings > Privacy
2. Enable Share runs with public Actor creators

This lets us see your run details when something goes wrong, so we can fix issues faster. Your data is only visible to the actor developer, not publicly.

Support

Found a bug or have a feature request? Open an issue in the Issues tab on this actor's page. For custom integrations — such as adding EFSA EU recall data, Canadian Food Inspection Agency records, or custom commodity coverage — reach out through the Apify platform.