UK Regulatory Ecosystem MCP

UK regulatory intelligence across 8 government databases — this MCP server gives AI agents a single interface to query Companies House, UK Police crime data, Food Standards Agency hygiene ratings, Land Registry, Charity Commission, Environment Agency waste carriers, flood warnings, and NSTA oil/gas licensing simultaneously. It is built for compliance teams, due diligence analysts, and AI workflows that need to assess UK entity risk without manually querying eight separate government portals.

The server orchestrates 8 Apify actors in parallel, constructs a cross-agency regulatory network graph, and runs 8 analysis tools ranging from KDE-based geographic risk heat mapping to multi-factor weighted risk scoring. Every tool call returns structured JSON with ranked risk scores, compliance flags, and actionable recommendations. Data is fetched live with a 5-minute in-memory cache to prevent redundant actor calls within a session.

What data can you access?

Why use UK Regulatory Ecosystem MCP?

Checking a UK company across all relevant regulatory databases manually requires opening eight different government websites, searching each one separately, and stitching results together by hand. For a single acquisition target, that typically takes 4–6 hours per analyst. For bulk screening — supply chains, charity registers, landlord portfolios — it scales to weeks.

This MCP server automates the entire process. One tool call triggers parallel queries across all selected data sources, builds a regulatory network graph linking entities by location and type, and runs whichever analysis algorithm the tool implements — compliance scoring, cross-agency violation detection, geographic heat mapping, or full report generation.

Beyond saving time, the cross-agency analysis surface risks that no single-source check would catch. A waste carrier with an expired registration that also operates near a crime hotspot and has a food hygiene rating of 1 presents compounding risk that only becomes visible when all three sources are queried together.

Platform benefits of running on Apify:

• Scheduling — run daily or weekly compliance sweeps on watchlists to catch status changes
• API access — trigger analysis from Python, JavaScript, or any HTTP client with a single token
• Spend controls — set a maximum per-run budget; the server stops charging when the limit is reached
• Monitoring — receive Slack or email alerts when runs fail or return unexpected results
• Integrations — connect output to Zapier, Make, Google Sheets, HubSpot, or downstream webhooks

Features

• 8 government data sources in one interface — Companies House, Police, FSA, Land Registry, Charity Commission, Environment Agency, Flood Warnings, and NSTA, all queryable from a single MCP tool call
• Parallel actor orchestration with 5-minute cache — the resolveData() function fans out up to 8 Apify actor calls concurrently using Promise.all(), with per-source TTL caching to avoid redundant charges within a session
• Regulatory network graph construction — buildRegNetwork() assembles entities from all sources into typed nodes (company, charity, food_business, waste_carrier, flood_zone, oil_license, crime_area, property) and creates co-location edges using 15-character location prefix matching
• KDE-based geographic risk heat mapping — computeGeographicRiskHeat() applies a Gaussian kernel (bandwidth = 0.5 degrees) to smooth crime, flood, and environmental risk signals across geographic cells, producing composite risk scores weighted 40% crime / 35% flood / 25% environmental
• Weighted multi-factor risk scoring — scoreMultiRegulatoryRisk() computes entity risk using 6 weighted factors: registry status (0.20), food hygiene (0.15), crime proximity (0.20), flood exposure (0.15), environmental compliance (0.15), network anomaly (0.15)
• Cross-agency violation detection — identifies entities simultaneously flagged by UK Police, Food Standards Agency, Environment Agency, and Flood Authority, classifying violations as dual_agency_flag or multi_agency_convergence
• Company compliance assessment — evaluates Companies House active status, linked food hygiene scores, waste carrier validity, and crime proximity into a 0–1 overall compliance score
• Charity legitimacy screening — detects revenue anomalies against population median, corporate director overlap by address, and registration status gaps, with scored red flag accumulation
• Environmental compliance analysis — assesses waste carrier active/expired status, oil/gas licence validity from NSTA, and flood zone edge proximity (0.3 risk increment per flood edge, capped at 1.0)
• Comprehensive regulatory report generation — generateRegulatoryReport() produces structured findings per category (Companies House, Food Standards, Environment Agency, Flood Risk, Police) with severity ratings (low / medium / high / critical) and specific escalation recommendations
• Per-tool pay-per-event pricing — 6 tools at $0.045/call, 2 tools (detect_cross_agency_violations and generate_regulatory_report) at $0.050/call; spending limit enforced server-side
• Dual transport support — serves both SSE (/sse + /messages) for legacy MCP clients and StreamableHTTP (/mcp) for modern clients, with a /health endpoint for readiness checks

Use cases for UK regulatory intelligence

Due diligence for UK acquisitions and investments

M&A analysts and private equity teams need to screen acquisition targets across all relevant UK registries before signing. A single generate_regulatory_report call against the target company name surfaces dissolved subsidiaries, environmental liabilities, food hygiene issues in operating premises, flood zone exposure across owned properties, and any cross-agency flags — all in one structured report, ready to paste into a due diligence memo.

Supply chain compliance monitoring

Procurement teams managing UK supplier networks need ongoing assurance that waste carriers hold valid Environment Agency registrations, food suppliers maintain acceptable FHRS ratings, and no supplier has newly dissolved a key operating entity. Scheduling a weekly assess_company_compliance sweep across a supplier watchlist provides continuous monitoring without manual portal checks.

Charity fraud and governance screening

Grant-making bodies, regulators, and investigative journalists need to verify that charities receiving public money are legitimately registered, have plausible income patterns, and lack suspicious director overlap with commercial entities. The screen_charity_legitimacy tool compares each charity's income against the population median, flags corporate co-location overlap, and produces a 0–1 legitimacy score with named red flags.

Geographic risk assessment for property investment

Real estate investors and urban planners evaluating UK locations need to understand the combined risk profile of an area — not just flood risk in isolation, but the intersection of crime density, environmental contamination from waste operators, and active flood warnings. The compute_geographic_risk_heat tool produces KDE-smoothed composite risk scores per geographic cell, enabling area-to-area comparison before committing capital.

Regulatory compliance AI agents

Compliance teams building AI agent workflows in Claude, Cursor, or other MCP-compatible clients can wire this server into their assistants to answer natural-language queries like "Is Meridian Waste Solutions compliant across all UK registries?" or "Which of our 50 suppliers are in high flood-risk zones?" The MCP protocol handles tool discovery and structured response formatting automatically.

Environmental enforcement support

Environment Agency regional teams, local authority environmental health officers, and NGOs monitoring illegal waste activity can use detect_cross_agency_violations to surface operators simultaneously appearing in crime data and waste carrier records with expired registrations — the combination that most commonly indicates fly-tipping or illegal waste treatment operations.

How to connect this MCP server to your AI client

Step 1: Get your Apify API token

Sign up at apify.com and copy your API token from Account Settings.

Step 2: Add to your MCP client

Claude Desktop — add to claude_desktop_config.json:

{
  "mcpServers": {
    "uk-regulatory-ecosystem": {
      "url": "https://uk-regulatory-ecosystem-mcp.apify.actor/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_TOKEN"
      }
    }
  }
}

Cursor / Windsurf / Cline — add the same URL and Authorization header in your MCP server settings panel.

Step 3: Start querying

Once connected, your AI client discovers all 8 tools automatically. Ask questions in natural language:

"Assess the regulatory compliance of Apex Industrial Services Ltd across all UK sources."

"Generate a regulatory report for businesses in the SW1A 1AA postcode area."

Step 4: Review structured results

Each tool returns a JSON object with ranked entities, risk scores, compliance flags, and recommendations. Results can be piped into a report, stored to a dataset, or used as context for downstream AI reasoning.

MCP tools

Tool input parameters

Every tool accepts the same two parameters:

Input examples

Company compliance check (default sources):

{
  "query": "Meridian Waste Solutions Ltd",
  "sources": ["companies", "police", "food", "waste"]
}

Geographic risk assessment for a postcode:

{
  "query": "E14 5AB",
  "sources": ["police", "flood", "waste", "energy"]
}

Full regulatory report with all 8 sources:

{
  "query": "Pinnacle Food Group",
  "sources": ["companies", "police", "food", "land", "charity", "waste", "flood", "energy"]
}

Input tips

• Use the company's registered name — search terms are passed directly to each source's API; the full legal name (as registered at Companies House) returns the most accurate results
• Start with 3–4 sources — the default source sets for each tool are optimised for that tool's purpose; expand to all 8 only when you need the comprehensive report
• Use postcode queries for geographic tools — compute_geographic_risk_heat and detect_cross_agency_violations return richer results when queried with a postcode or area name rather than a company name
• Combine tools in sequence — run score_multi_regulatory_risk first to identify the highest-risk entities, then call generate_regulatory_report only on those flagged above 0.5

Output examples

assess_company_compliance output

{
  "checks": [
    {
      "company": "Pinnacle Food Group Ltd",
      "registryStatus": "Active",
      "foodHygieneScore": 2.4,
      "wasteCarrierValid": false,
      "crimeProximity": 0.4,
      "overallCompliance": 0.3800
    },
    {
      "company": "Meridian Catering Services",
      "registryStatus": "Active",
      "foodHygieneScore": 4.2,
      "wasteCarrierValid": true,
      "crimeProximity": 0.0,
      "overallCompliance": 0.8600
    }
  ],
  "totalChecked": 2,
  "failedCount": 1,
  "avgCompliance": 0.6200
}

detect_cross_agency_violations output

{
  "violations": [
    {
      "entity": "Apex Clearance Ltd",
      "agencies": ["UK Police", "Environment Agency", "Food Standards Agency"],
      "violationType": "multi_agency_convergence",
      "severity": 0.7500,
      "evidence": "Entity flagged across 3 regulatory agencies"
    }
  ],
  "totalDetected": 1,
  "multiAgencyCount": 1,
  "avgSeverity": 0.7500
}

generate_regulatory_report output

{
  "findings": [
    {
      "category": "Companies House",
      "finding": "2 dissolved/liquidated company(ies) with regulatory links",
      "severity": "high",
      "evidence": "Apex Waste UK, Meridian En",
      "recommendation": "Report to Companies House and investigate potential phoenix company fraud"
    },
    {
      "category": "Food Standards",
      "finding": "3 food business(es) with hygiene rating <= 2",
      "severity": "high",
      "evidence": "Harbour Kitchen (1), Eastgate Cafe (2), The Diner (2)",
      "recommendation": "Escalate to local authority food safety team for urgent inspection"
    },
    {
      "category": "Environment Agency",
      "finding": "1 waste carrier(s) with non-active registration",
      "severity": "high",
      "evidence": "Apex Clearance Ltd",
      "recommendation": "Report to Environment Agency for illegal waste carriage investigation"
    }
  ],
  "overallRisk": "high",
  "entitiesScreened": 24,
  "recommendations": [
    "Companies House referral for dissolved entity activity",
    "Food Standards Agency urgent inspection programme",
    "Environment Agency enforcement action for unlicensed waste carriers"
  ]
}

compute_geographic_risk_heat output

{
  "cells": [
    {
      "area": "Silvertown Way, E16",
      "lat": 51.5074,
      "lng": 0.0123,
      "crimeRisk": 0.7200,
      "floodRisk": 0.8100,
      "environmentalRisk": 0.4500,
      "compositeRisk": 0.6955
    },
    {
      "area": "Thames Road, DA8",
      "lat": 51.4892,
      "lng": 0.1834,
      "crimeRisk": 0.3100,
      "floodRisk": 0.6400,
      "environmentalRisk": 0.2800,
      "compositeRisk": 0.4700
    }
  ],
  "totalAreas": 14,
  "highRiskCount": 3,
  "avgCompositeRisk": 0.5210
}

Output fields

assess_company_compliance

detect_cross_agency_violations

compute_geographic_risk_heat

generate_regulatory_report

score_multi_regulatory_risk

How much does it cost to run UK regulatory intelligence queries?

This MCP uses pay-per-event pricing — you pay per tool call. Six tools cost $0.045 per call, and two tools (detect_cross_agency_violations and generate_regulatory_report) cost $0.050 per call. Platform compute costs are included.

You can set a maximum spending limit per run to control costs. The server returns a spending-limit-reached message and stops charging when your budget is reached.

The Apify Free plan includes $5 of monthly platform credits, covering approximately 110 tool calls. Most individual due diligence workflows cost under $0.50. Compare this to commercial compliance platforms that charge $500–2,000/month for similar UK registry coverage — with this MCP, most teams spend $5–30/month with no subscription commitment.

Using UK regulatory intelligence via the API

Python

from apify_client import ApifyClient

client = ApifyClient("YOUR_API_TOKEN")

run = client.actor("ryanclinton/uk-regulatory-ecosystem-mcp").call(run_input={})

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(f"Tool: {item.get('tool')} | Result: {item.get('result')}")

JavaScript

import { ApifyClient } from "apify-client";

const client = new ApifyClient({ token: "YOUR_API_TOKEN" });

const run = await client.actor("ryanclinton/uk-regulatory-ecosystem-mcp").call({});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
for (const item of items) {
    console.log(`Tool: ${item.tool} | Overall Risk: ${item.result?.overallRisk}`);
}

cURL (direct MCP protocol — Streamable HTTP)

# Call the assess_company_compliance tool
curl -X POST "https://uk-regulatory-ecosystem-mcp.apify.actor/mcp" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_APIFY_TOKEN" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
      "name": "assess_company_compliance",
      "arguments": {
        "query": "Apex Industrial Services Ltd",
        "sources": ["companies", "police", "food", "waste"]
      }
    },
    "id": 1
  }'

# Call generate_regulatory_report for a full multi-source report
curl -X POST "https://uk-regulatory-ecosystem-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_regulatory_report",
      "arguments": {
        "query": "Pinnacle Waste Group",
        "sources": ["companies", "police", "food", "land", "charity", "waste", "flood", "energy"]
      }
    },
    "id": 2
  }'

How UK Regulatory Ecosystem MCP works

Phase 1: Parallel actor fan-out with caching

When a tool is called, resolveData() fans out up to 8 concurrent Apify actor calls using Promise.all(). Each source maps to a dedicated Apify actor: ryanclinton/uk-companies-house, ryanclinton/uk-police-crime-data, ryanclinton/uk-food-hygiene, ryanclinton/uk-land-registry, ryanclinton/uk-charity-commission, ryanclinton/ea-waste-carriers, ryanclinton/uk-flood-warnings, and ryanclinton/nsta-oil-gas-licenses. Each actor is called with waitSecs: 180 and memory: 256MB, returning up to 500 items per source. Results are cached in a module-level Map with 5-minute TTL, keyed by {source}:{query}, so repeated tool calls within a session do not re-run actors.

Phase 2: Regulatory network graph construction

buildRegNetwork() converts raw actor output arrays into a typed graph. Each data source produces nodes of a specific type: companies produce company nodes with Companies House status metadata, police data produces crime_area nodes with latitude/longitude and crime category, food hygiene produces food_business nodes with FHRS rating, land registry produces property nodes with tenure and price paid, charities produce charity nodes with income and registration status, waste carriers produce waste_carrier nodes with expiry date, flood warnings produce flood_zone nodes with severity level, and oil/gas data produces oil_license nodes with NSTA operator metadata.

Edges are created by two mechanisms: co-location matching (15-character prefix comparison of location strings across different node types, creating operates_at edges) and flood risk linking (8-character location prefix matching between flood zones and properties/companies/food businesses, creating flood_risk edges with weight 0.7).

Phase 3: Analysis algorithm execution

Each tool applies a different algorithm to the network. assessCompanyCompliance() traverses company node edges to find linked food, waste, and crime nodes, computing a weighted composite score (registry 0.3, food hygiene 0.3, waste 0.2, crime inverse 0.2). detectCrossAgencyViolations() identifies company nodes connected to crime, low-hygiene food, inactive waste, and flood nodes simultaneously, classifying by agency count. computeGeographicRiskHeat() groups all nodes by area prefix, accumulates risk contributions per type (crime +0.2, flood +0.3, waste/oil +0.15), then applies a Gaussian kernel with bandwidth 0.5 degrees to smooth values across nearby cells, producing composite scores weighted 40/35/25. scoreMultiRegulatoryRisk() uses a six-factor model with explicit weights: registry status (0.20), food hygiene (0.15), crime proximity (0.20), flood exposure (0.15), environmental (0.15), network anomaly (0.15). generateRegulatoryReport() applies category-specific rules to detect dissolved companies with active links, food ratings ≤ 2, inactive waste carriers, active flood zones, and crime hotspots above threshold count.

Phase 4: Transport and response

The server runs on Express with two MCP transport endpoints: GET /sse + POST /messages (SSE protocol, legacy client compatibility) and POST /mcp (StreamableHTTP, modern clients). In Apify Standby mode it binds to process.env.PORT (default 3031) and runs persistently. In normal actor run mode it binds to port 8080, logs a status message, and exits after 1 second — ensuring the actor does not hang when triggered outside Standby context.

Tips for best results

1. Use the most focused source set for each tool. The default source sets are designed for each tool's analysis logic. Using all 8 sources for a screen_charity_legitimacy call adds cost and latency without improving the charity scoring algorithm, which only uses charity and companies nodes.

2. Query by registered company name for Companies House lookups. The query string is passed directly to each actor. Using the exact legal name as registered at Companies House (e.g., "Apex Industrial Services Limited" not "Apex Industrial") returns more precise results from the Companies House actor.

3. Run score_multi_regulatory_risk as a triage step. Its six-factor model covers the full source set efficiently. Use the resulting riskLevel classification to decide which entities warrant a full generate_regulatory_report call — only escalating entities above medium reduces cost significantly.

4. Use postcode queries for geographic tools. compute_geographic_risk_heat and detect_cross_agency_violations produce richer results when queried with a UK postcode or borough name, because geographic tools rely on location-string matching across crime, flood, and environmental nodes.

5. Set a spending limit for bulk sweeps. When processing a large supplier list, set a maximum spend per MCP session. The server enforces the limit server-side and returns a structured error rather than silently stopping — your agent can handle this as a pagination signal.

6. Chain tools for richer reports. Call map_regulatory_network first to understand the full entity graph and network density, then call detect_cross_agency_violations on the highest-degree nodes. The network map output guides which follow-up tools are most valuable.

7. For charity due diligence, always include companies in sources. The screen_charity_legitimacy algorithm's director overlap detection works by matching charity and company addresses. Without the companies source, overlap detection returns zero regardless of actual relationships.

Combine with other Apify actors

Limitations

• Location matching is prefix-based, not geocoded. The network graph's edge construction uses 15-character and 8-character string prefix matching on location fields. Entities in the same area but with differently formatted address strings may not be connected. For precise spatial analysis, geocoded matching would be required.
• Up to 500 items per source per query. Each Apify actor call returns a maximum of 500 items. For queries matching large result sets (major UK cities, common company names), the network graph reflects a sample rather than the complete universe.
• The server must run in Apify Standby mode for persistent MCP connections. When triggered as a standard actor run (not via the MCP URL), the server starts, logs a status message, and exits after 1 second. It is not designed for one-off batch processing jobs.
• Scotland and Northern Ireland have partial coverage. Food Standards Agency (FHRS) coverage is England, Wales, and Northern Ireland; Scotland uses Food Standards Scotland with a different rating system. Police crime data coverage varies by force — not all Scottish forces publish to the national API.
• Real-time flood warnings reflect current status only. Flood warning data is live at query time. Historical flood event data is not included. For flood risk assessment of specific properties, check Environment Agency flood map records separately.
• Cross-agency violation detection requires collocated entities. The algorithm detects violations by traversing network edges. If a company and its food premises have different address strings (e.g., registered office vs. trading address), they may not be connected in the graph, reducing detection accuracy.
• Oil/gas data covers NSTA-licensed North Sea operations. The energy source (nstaOilGas) covers PEDL and offshore licences from the North Sea Transition Authority. It does not cover onshore renewable energy, water abstraction licences, or other energy-related environmental permits.
• Charity income comparison is within-query population only. Revenue anomaly detection compares each charity's income to the median of charities returned by the same query. A query returning only one charity produces a median of that charity's own income, making anomaly detection meaningless. Query multiple charities together for reliable anomaly scoring.

Integrations

• Apify API — trigger UK regulatory analysis programmatically from Python, JavaScript, or any HTTP client and store results in Apify datasets
• Zapier — build no-code workflows that trigger a compliance sweep when a new supplier is added to a spreadsheet and alert the team on critical findings
• Make — automate charity screening workflows triggered on a schedule or by webhook, routing high-risk results to an email or Slack channel
• Google Sheets — export risk scores and compliance flags for supplier watchlists directly into a shared spreadsheet for team review
• Webhooks — fire alerts to Slack, Teams, or a compliance dashboard whenever a scheduled run produces entities with riskLevel: critical
• LangChain / LlamaIndex — connect this MCP server as a tool in LangChain agent chains or LlamaIndex workflows for multi-step UK regulatory research pipelines

Troubleshooting

No entities returned despite a known company existing in Companies House. The search term is passed directly to the underlying actor. Ensure the query matches the company's full registered name, including "Limited", "Ltd", or "PLC". Partial name matches and trading names may return no results from the Companies House actor.

Risk scores appear uniform at 0.5 or similar round numbers. This occurs when the query returns data from only one source with no co-location matches, producing a network with nodes but no edges. Without edges, compliance and risk algorithms fall back to default values (food hygiene defaults to 0.5, crime proximity to 0). Add more sources to the sources array or use a more specific query to improve entity matching across sources.

The server returns a spending-limit-reached message and stops. This is expected behaviour. Set a higher maxTotalChargeUsd limit in your Apify account or split the analysis across multiple smaller queries. The server enforces the limit per event, so partial results from tools called before the limit was hit are still returned.

Cross-agency violations not detected despite known issues. Violation detection requires at least 2 different agency node types to be connected to the same company node via edges. If address strings differ between sources (e.g., "14 High Street" vs. "14 High St."), the edge will not be created. Try simplifying the query to a postcode or borough name to increase co-location matches.

SSE connection drops after a few minutes. Some reverse proxies and cloud load balancers terminate idle SSE connections. Switch to the StreamableHTTP endpoint (/mcp) if your client supports it, or configure your proxy to increase the idle connection timeout.

Responsible use

• All data sources queried by this MCP server are publicly available UK government registries published under Open Government Licence or equivalent.
• Use of Companies House data is subject to the Companies House terms of use.
• Comply with UK GDPR and the Data Protection Act 2018 when storing or processing personal data derived from regulatory records.
• Do not use regulatory risk scores as the sole basis for decisions that could materially affect individuals or organisations without human review and corroboration from authoritative sources.
• For guidance on web scraping and data access legality, see Apify's guide.

FAQ

What UK regulatory databases does this MCP server cover? Eight databases: Companies House (company registrations, officers, filings, PSC), Food Standards Agency (FHRS hygiene ratings 0–5), UK Police API (street-level crime and outcomes), UK Land Registry (ownership, title, price paid), Charity Commission (registration, income, governance), Environment Agency (waste carrier registrations), UK Flood Warnings API (active alerts and zones), and NSTA (North Sea Transition Authority oil/gas licences).

How many sources can I query in a single tool call? All 8 simultaneously. The generate_regulatory_report and score_multi_regulatory_risk tools default to all 8. You can pass any subset via the sources parameter. The server fans out actor calls in parallel, so 8 sources does not take 8× as long as 1.

How current is the data returned? Data is fetched live from each source at query time with a 5-minute in-memory cache. If you call the same tool with the same query twice within 5 minutes, the second call uses cached results at no additional charge. Outside that window, fresh actor runs are triggered.

Does this MCP server cover Scottish and Northern Irish entities? Companies House covers all UK-incorporated entities. Food Standards coverage is England, Wales, and Northern Ireland (Scotland uses a different system). Police crime data coverage varies by force — Scottish forces use separate systems with limited API exposure. For Scottish-specific regulatory intelligence, supplement with dedicated Scottish sources.

How does the geographic risk heat map work? compute_geographic_risk_heat groups all nodes by location prefix, accumulates per-type risk contributions (crime +0.2, flood +0.3, waste/oil +0.15), then applies a Gaussian kernel with bandwidth 0.5 degrees to smooth values across nearby cells. The composite risk score weights crime 40%, flood 35%, and environmental 25%. Cells with compositeRisk > 0.6 are classified as high risk.

How is the multi-regulatory risk score calculated? score_multi_regulatory_risk uses six weighted factors: registry status (weight 0.20), food hygiene (0.15), crime proximity (0.20), flood exposure (0.15), environmental compliance (0.15), and network anomaly (0.15). Each factor produces a 0–1 score. The weighted sum maps to low (<0.25), medium (<0.50), high (<0.75), or critical (≥0.75).

Is it legal to query these UK government databases? Yes. All 8 data sources are publicly available UK government registries published under Open Government Licence or comparable terms. See Apify's guide on web scraping legality for a full discussion.

How is this different from a manual Companies House search? A Companies House search gives you one company's filing history. This MCP simultaneously queries 8 databases, builds a cross-agency network graph linking the company to its food premises, waste carrier registrations, nearby crime incidents, flood zones, and charity connections, then computes a composite risk score across all of them. It surfaces risk patterns that no single-source search could detect.

Can I use this for bulk supplier screening? Yes. Call assess_company_compliance or score_multi_regulatory_risk for each supplier name in your list. The 5-minute cache means repeated queries within a session are free. For lists of 50+ suppliers, set a per-session spending limit and process in batches to control cost.

How long does a typical tool call take? Response time depends on how many sources are queried. A 2-source query typically completes in 20–40 seconds. A full 8-source generate_regulatory_report call may take 60–120 seconds, as each actor call has a 180-second timeout. The parallel fan-out means total time is closer to the slowest single actor rather than the sum of all actors.

Can I schedule this MCP for recurring compliance monitoring? You can schedule actor runs via the Apify scheduler, but for MCP usage the server is designed for on-demand queries from connected clients. For scheduled compliance sweeps, consider triggering the underlying Apify actors directly (e.g., ryanclinton/uk-companies-house) on a cron schedule and processing results with a downstream workflow.

What happens if one of the 8 underlying actors fails? The runActor() function wraps each actor call in a try-catch. If an actor fails or times out, it returns an empty array and logs an error. The network graph is built from the remaining successful sources. The tool returns results based on whatever data was successfully retrieved, and the failing source is effectively excluded from that call's analysis.

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 solutions or enterprise integrations, reach out through the Apify platform.