Higher Education Risk MCP Server

Higher education risk intelligence for any university, college, or for-profit institution — delivered through the Model Context Protocol. This MCP server connects your AI assistant to 8 live data sources and returns structured risk assessments covering accreditation standing, research integrity, federal funding dependency, and regulatory exposure. Purpose-built for accreditors, student loan portfolio managers, state regulators, and institutional researchers who need objective institution health data without manual database searching.

The server runs in persistent standby mode on the Apify platform, accepting tool calls from Claude Desktop, Cursor, Windsurf, Cline, or any MCP-compatible client. Each request fans out to multiple data sources in parallel — CFPB complaints, Federal Register publications, OpenAlex publications, ORCID researcher profiles, USAspending awards, Grants.gov portfolios, and Wikipedia — then applies four scoring models to produce a composite risk score with plain-language signals.

What data can you extract?

Why use Higher Education Risk MCP Server?

Manual institution due diligence means switching between the CFPB complaint portal, Federal Register search, USAspending.gov, Grants.gov, OpenAlex, and ORCID — then building spreadsheets to compare results. A single institution screening can take 4-6 hours. For a portfolio of 20 schools, that is weeks of analyst time.

This MCP server automates the entire process. Call one tool, get a structured risk assessment in under 2 minutes. Run peer comparisons across 5 institutions simultaneously. Schedule quarterly monitoring without touching a spreadsheet.

• Scheduling — run institution portfolio sweeps on a quarterly or annual cadence to track risk trajectory over time
• API access — trigger assessments from Python, JavaScript, or any HTTP client for programmatic integration into risk platforms
• Parallel data collection — up to 8 data sources queried simultaneously per tool call, not sequentially
• Monitoring — get Slack or email alerts when institution assessments complete or when spending limits are approached
• Integrations — connect to Zapier, Make, Google Sheets, HubSpot, or any webhook-capable system for post-processing

Features

• Four independent scoring models — Accreditation Risk (0-100, risk-oriented), Research Integrity Index (0-100, quality-oriented), Financial Viability Score (0-100, health-oriented), and Regulatory Exposure Score (0-100, risk-oriented) computed from separate data source combinations
• Composite risk score — a single 0-100 institution health number derived as a weighted average of all four models: (100 − accreditation) × 0.25 + research × 0.25 + financial × 0.25 + (100 − regulatory) × 0.25
• Five-tier risk classification — LOW, MODERATE, ELEVATED, HIGH, and CRITICAL labels mapped to composite score bands for fast triage
• Complaint acceleration detection — compares complaint volume in the trailing 6 months against the prior period; flags when recent count exceeds 1.5× older count, signalling a deteriorating trajectory
• Title IV threat scoring — Federal Register documents are scanned for "Title IV", "student aid", "gainful employment", "borrower defense", and "accreditation" keywords, with each hit weighted toward the regulatory exposure score
• For-profit school pattern detection — specifically scores elevated risk when Wikipedia or deep research content contains both "for-profit" and "criticism" or "lawsuit"
• Citation health scoring — uses log2(avgCitations) × 5 to produce a logarithmic citation quality score, avoiding inflation from a single high-cited outlier paper
• Journal diversity scoring — counts unique publication venues across OpenAlex results; institutions publishing across 5+ journals receive a diversity bonus
• Closure risk projection (1-5 scale) — converts financial viability score to a 5-point closure risk rating aligned to common institutional oversight frameworks
• Strengths and risks narrative — the full dossier tool auto-generates plain-language strength and risk bullet points from scoring signals for direct use in reports
• Parallel actor orchestration — runActorsParallel dispatches all data source calls via Promise.allSettled, so individual source failures do not abort the entire assessment
• Spending limit guards — every tool checks Actor.charge() result and returns a clean error message if the per-run event charge limit has been reached
• 8 MCP tools covering targeted and comprehensive use cases from quick snapshots to full dossiers

Use cases for higher education risk intelligence

Accreditation agency risk-based monitoring

Regional and national accreditors carry oversight responsibility for hundreds of institutions. Manually tracking complaint trends, regulatory actions, and financial signals across a large portfolio is prohibitive. This server gives accreditation staff a structured score for every institution at review time — surfacing which schools need deeper scrutiny and which are stable — so oversight resources focus where the risk is highest.

Student loan portfolio management

Servicers and investors holding student loan portfolios need early warning when institutions show closure risk indicators. Institutional closures trigger borrower defense claims and loan discharges. The Financial Viability Model and Accreditation Risk Score provide leading indicators — declining federal funding, rising complaint volumes, and Federal Register enforcement patterns — before a closure event becomes public.

State higher education regulatory oversight

State higher education agencies monitor dozens to hundreds of institutions for consumer protection, degree integrity, and financial stability. This server's compare_peer_institutions and scan_student_complaints tools let state staff quickly identify which institutions in their jurisdiction are accumulating student complaints or regulatory flags without running separate database searches.

Pre-enrollment due diligence for prospective students

Students and families committing to multi-year enrollment and student loan obligations benefit from objective institutional health data. The assess_institution_risk tool returns a plain-language institutional snapshot — complaint volume, accreditation signals, research credentials — in seconds, supporting informed enrollment decisions.

Institutional investor and private equity due diligence

Education sector investors evaluating acquisitions of for-profit institutions need financial viability data, complaint exposure, and regulatory risk profiles alongside traditional financial due diligence. This server surfaces the public-data risk signals that inform pricing and deal structure for higher education transactions.

Academic research and policy analysis

Higher education researchers studying institutional quality, closure patterns, or regulatory effectiveness can use this server to build longitudinal datasets across hundreds of institutions. The structured JSON output integrates cleanly with Python analysis pipelines, reducing the manual data collection burden for large-scale studies.

How to use the Higher Education Risk MCP Server

Connect to Claude Desktop

Add this entry to your claude_desktop_config.json:

{
  "mcpServers": {
    "higher-education-risk": {
      "url": "https://higher-education-risk-mcp.apify.actor/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_TOKEN"
      }
    }
  }
}

Restart Claude Desktop. The 8 higher education risk tools will appear in your tools panel.

Run your first assessment

1. Name the institution — type the full institution name as it appears officially: "University of Phoenix", "DeVry University", "Western Governors University", or any accredited institution
2. Choose the right tool — use assess_institution_risk for a quick snapshot, generate_institution_dossier for comprehensive due diligence, or a specialized tool for targeted analysis
3. Wait for data collection — the server queries up to 8 data sources in parallel; most tools complete in 60-90 seconds
4. Read the structured result — the response includes a risk score, classification level, plain-language signals, and raw supporting data you can use directly in reports

MCP tools

Output example

The following is representative output from generate_institution_dossier for a mid-sized for-profit institution:

{
  "institution": "Pinnacle Career Institute",
  "compositeScore": 38,
  "overallRisk": "ELEVATED",
  "accreditationRisk": {
    "score": 61,
    "complaintVolume": 34,
    "regulatoryActions": 18,
    "reputationFlags": 14,
    "riskLevel": "AT_RISK",
    "signals": [
      "34 consumer complaints — elevated student dissatisfaction",
      "Multiple regulatory actions in Federal Register — accreditation under scrutiny",
      "Wikipedia/research sources flag institutional controversies or closure risk",
      "Complaint acceleration: 19 in last 6 months vs 9 prior — deteriorating trajectory"
    ]
  },
  "researchIntegrity": {
    "score": 29,
    "totalPapers": 4,
    "totalResearchers": 2,
    "citationHealth": 3,
    "integrityLevel": "BELOW_AVERAGE",
    "signals": [
      "Minimal research output — limited academic credibility",
      "No ORCID-registered researchers found — research credibility concern"
    ]
  },
  "financialViability": {
    "score": 34,
    "federalFunding": 2800000,
    "grantCount": 1,
    "fundingDiversity": 1,
    "closureRisk": 2,
    "signals": [
      "Low federal funding level — financial vulnerability",
      "No active grants found — limited competitive funding success"
    ]
  },
  "regulatoryExposure": {
    "score": 58,
    "activeRules": 11,
    "titleIVThreats": 4,
    "enforcementActions": 2,
    "exposureLevel": "HIGH",
    "signals": [
      "4 Title IV / student aid regulatory actions — federal funding at risk",
      "2 enforcement actions — active regulatory scrutiny",
      "12 student loan/education complaints — regulatory trigger risk"
    ]
  },
  "allSignals": [
    "34 consumer complaints — elevated student dissatisfaction",
    "Multiple regulatory actions in Federal Register — accreditation under scrutiny",
    "Complaint acceleration: 19 in last 6 months vs 9 prior — deteriorating trajectory",
    "Minimal research output — limited academic credibility",
    "Low federal funding level — financial vulnerability",
    "4 Title IV / student aid regulatory actions — federal funding at risk"
  ],
  "strengths": [],
  "risks": [
    "Accreditation AT_RISK — 34 complaints, 18 regulatory actions",
    "Closure risk elevated (2/5) — limited federal funding and grant activity",
    "HIGH regulatory exposure — 4 Title IV threats",
    "Poor research integrity — minimal publications and no ORCID researchers"
  ]
}

Output fields

How much does it cost to assess higher education risk?

This server uses pay-per-event pricing — you pay $0.045 per tool call. Platform compute costs are included. There is no subscription fee. The Apify Free plan includes $5 of monthly credits, which covers over 100 institution assessments.

You can set a maximum spending limit per run to control costs. The server returns a clean error message if the per-run event charge limit is reached, and the run stops gracefully.

Compare this to manual database searches across CFPB, Federal Register, USAspending, OpenAlex, and ORCID — averaging 4-6 hours per institution at analyst rates of $50-150/hour. A 50-institution portfolio sweep that costs $2.25 here would cost $10,000-30,000 in analyst time.

Using the Higher Education Risk MCP Server via the API

Python

from apify_client import ApifyClient

client = ApifyClient("YOUR_API_TOKEN")

run = client.actor("ryanclinton/higher-education-risk-mcp").call(run_input={})

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(f"Institution: {item.get('institution')} | Risk: {item.get('overallRisk')} | Score: {item.get('compositeScore')}")

For direct MCP tool calls from Python, use the HTTP endpoint:

import requests

response = requests.post(
    "https://higher-education-risk-mcp.apify.actor/mcp",
    headers={
        "Content-Type": "application/json",
        "Authorization": "Bearer YOUR_APIFY_TOKEN"
    },
    json={
        "jsonrpc": "2.0",
        "method": "tools/call",
        "params": {
            "name": "generate_institution_dossier",
            "arguments": {"institution": "Grand Canyon University"}
        },
        "id": 1
    }
)
dossier = response.json()
result = dossier["result"]["content"][0]["text"]
print(result)

JavaScript

import { ApifyClient } from "apify-client";

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

// Direct MCP HTTP call
const response = await fetch("https://higher-education-risk-mcp.apify.actor/mcp", {
    method: "POST",
    headers: {
        "Content-Type": "application/json",
        "Authorization": "Bearer YOUR_API_TOKEN"
    },
    body: JSON.stringify({
        jsonrpc: "2.0",
        method: "tools/call",
        params: {
            name: "compare_peer_institutions",
            arguments: {
                institutions: [
                    "University of Phoenix",
                    "Western Governors University",
                    "Southern New Hampshire University"
                ]
            }
        },
        id: 1
    })
});

const data = await response.json();
const comparison = JSON.parse(data.result.content[0].text);
for (const inst of comparison.peerComparison) {
    console.log(`${inst.institution}: research=${inst.researchScore}, funding=${inst.fundingScore}, complaints=${inst.complaints}`);
}

cURL

# Full institution dossier
curl -X POST "https://higher-education-risk-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_institution_dossier",
      "arguments": {"institution": "DeVry University"}
    },
    "id": 1
  }'

# Quick accreditation exposure check
curl -X POST "https://higher-education-risk-mcp.apify.actor/mcp" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_APIFY_TOKEN" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {
      "name": "check_accreditation_exposure",
      "arguments": {"institution": "Corinthian Colleges"}
    },
    "id": 2
  }'

How Higher Education Risk MCP Server works

Data collection phase

Each tool call dispatches parallel requests to between 1 and 8 Apify actors using Promise.allSettled. Individual actor failures do not abort the assessment — the orchestrator records empty arrays for failed sources and proceeds with the data that arrived. The full dossier tool fans out to all 8 actors simultaneously: company-deep-research, openalex-research-papers, orcid-researcher-search, cfpb-consumer-complaints, federal-register-search, grants-gov-search, usaspending-search, and wikipedia-article-search. Each actor runs with 256 MB memory and a 120-second timeout.

Scoring model phase

Four independent scoring functions process the collected data:

Accreditation Risk (0-100, risk) combines complaint severity scoring (loan/fraud keywords weighted 3-5× ordinary complaints, max 30 points), Federal Register regulatory action density (accreditation/Title IV/sanction/closure keywords, max 25 points), reputation flags from Wikipedia and deep research content (closure/controversy/for-profit lawsuit patterns, max 25 points), and a complaint acceleration component that penalises institutions where the trailing 6-month complaint count exceeds 1.5× the prior period (max 20 points).

Research Integrity (0-100, quality) scores publication volume (3 points per paper, max 30), citation quality via log2(max(1, avgCitations)) × 5 (max 25), ORCID researcher count (3 points per researcher, max 25), and journal diversity (3 points per unique venue, max 20).

Financial Viability (0-100, health) maps total federal funding from USAspending to a tiered score ($100M+ = 30, $10M+ = 22, $1M+ = 15, $100K+ = 8, below = 2), adds grant portfolio diversity (5 points per funding agency, 2 per grant, max 25), federal award activity count (2 per award, max 25), and grant competitiveness (4 per Grants.gov listing, max 20).

Regulatory Exposure (0-100, risk) penalises Title IV/student aid document hits (6 points each), enforcement-type documents (4 points each), student-related CFPB complaints (3 points each), and Federal Register volume (2 points per entry).

Composite score assembly

The generateInstitutionDossier function inverts the two risk-oriented scores and averages all four at equal 25% weight: (100 − accreditation) × 0.25 + research × 0.25 + financial × 0.25 + (100 − regulatory) × 0.25. The result is a 0-100 health index where higher numbers indicate healthier institutions, mapped to five risk tiers (LOW ≥ 80, MODERATE ≥ 60, ELEVATED ≥ 40, HIGH ≥ 20, CRITICAL below 20).

MCP transport

The server uses StreamableHTTPServerTransport from the Model Context Protocol SDK with per-request server instantiation. It runs in Apify's standby mode, keeping the process alive between requests. The /mcp endpoint accepts POST only; GET and DELETE return 405. Non-standby runs exit after a 1-second health check to keep the Apify dataset clean.

Tips for best results

1. Use the full institution name. "University of Phoenix" returns better results than "UoP" or "Phoenix Uni". The actor queries pass directly to each data source API, and abbreviated names often return insufficient data, lowering scores artificially.

2. Run scan_student_complaints first for for-profit schools. CFPB complaint data is the most distinctive risk signal for for-profit institutions. Use this tool as a quick pre-screen before committing to a full dossier.

3. Treat closure risk 1-2 as a hard flag. The Financial Viability Model is calibrated conservatively. Closure risk ratings of 1 or 2 indicate genuinely thin federal funding and grant activity — conditions present before multiple well-known for-profit closures.

4. Interpret research integrity score in context. Community colleges and teaching-focused institutions will naturally score lower on research integrity because they do not produce research output. Low research scores are only concerning when paired with high accreditation risk.

5. Use compare_peer_institutions to validate scores. An accreditation risk of 55 means different things for a community college versus a research university. Comparing against 2-3 peer institutions of the same type provides calibration.

6. Schedule quarterly runs for portfolio monitoring. Complaint acceleration and Federal Register action patterns change over time. Quarterly monitoring catches deterioration before it becomes a closure event. Use Apify scheduling with a webhook to push results into your risk platform.

7. Pair with Company Deep Research for enrollment trend data. The deep research actor returns news and web content about enrollment trends and leadership changes that supplement the structured scoring.

Combine with other Apify actors

Limitations

• Data reflects public records only. Accreditation probation, sanction, or loss decisions not yet published to Federal Register or Wikipedia will not appear in scores. Private accreditor communications are not accessible.
• CFPB complaints are reported complaints, not verified ones. High complaint volume is a risk signal, not proof of wrongdoing. Some institutions with high enrollment naturally attract more complaints by volume.
• Research integrity scores are output-based, not process-based. The score measures publication volume and citation counts, not research ethics, IRB compliance, or retraction history. An institution can score GOOD on research integrity while having unreported misconduct.
• Financial Viability is a federal-data view only. The model uses USAspending and Grants.gov — federal funding sources. Tuition revenue, endowment size, bond ratings, and state appropriations are not available in this data and are significant factors in actual financial health.
• Federal Register search coverage depends on keyword matching. Institutions with names that appear in many regulatory contexts may produce noisy results. Institutions with uncommon names may miss relevant documents if they are referenced differently in federal publications.
• Wikipedia coverage is uneven. Large research universities have extensive Wikipedia entries. Small community colleges or recently-founded institutions may have minimal or no Wikipedia presence, which reduces the reputation scoring component's data quality.
• No real-time accreditor database access. The server does not connect to SACSCOC, HLC, ACICS, DEAC, or other accreditor portals. Accreditation status changes must be inferred from Federal Register and reputation sources.
• compare_peer_institutions runs institutions sequentially, not in parallel, to avoid resource contention. Comparing 5 institutions takes approximately 5× the time of a single comparison.

Integrations

• Apify API — call generate_institution_dossier from any HTTP client for programmatic integration into risk management platforms or regulatory portals
• Zapier — trigger institution assessments from a new row in a Google Sheet of schools under review, push results to Slack or a risk dashboard
• Make — build automated monitoring workflows that run quarterly institution sweeps and post flagged results to a compliance team channel
• Google Sheets — export composite scores and risk classifications for a portfolio of institutions into a tracking spreadsheet
• Webhooks — receive notification when a scheduled institution assessment completes, triggering downstream alerts in your risk management system
• LangChain / LlamaIndex — use the MCP server as a tool node in an AI research pipeline that synthesizes institution risk data with other intelligence sources

Troubleshooting

• Low scores despite a large university being assessed. Search quality varies by institution name. If querying "Penn State", try "Pennsylvania State University" instead. The actor queries pass institution names directly to each data source API, which performs its own text matching. More specific names generally return higher-quality results.

• Research integrity scores of POOR for known research universities. OpenAlex and ORCID queries are limited by the search API's matching precision. If a research-intensive institution scores unexpectedly low, try analyze_research_integrity with a specific department parameter (e.g., department: "engineering") to target a specific field and verify the underlying data is returning correctly.

• generate_institution_dossier taking longer than 3 minutes. The full dossier fans out to 8 actors. If the Apify platform is under load or a specific data source is slow, the 120-second actor timeout may delay individual sub-queries. The Promise.allSettled pattern ensures the dossier still returns — with empty arrays for any timed-out sources — rather than failing entirely.

• Financial viability scores appear low for well-funded private universities. Private universities often receive limited direct federal contracts and grants visible in USAspending because they receive federal funds through different mechanisms (Title IV disbursements go to students, not the institution directly). For private nonprofits, supplement with the analyze_research_integrity tool as a proxy for institutional research health.

• Spending limit reached error. If you see "eventChargeLimitReached": true, your Apify run has hit the per-run spending limit you configured. Increase the limit in your actor run options, or split your institution list across multiple runs.

Responsible use

• This server accesses only publicly available government databases and open academic registries.
• Risk scores are analytical models based on public data signals, not official accreditation status determinations.
• Do not use automated scores as the sole basis for consequential decisions such as denying students enrollment or revoking institutional approvals — human review is required.
• Comply with applicable data protection and privacy regulations when incorporating assessment outputs into institutional records.
• For guidance on the legal aspects of automated data collection, see Apify's guide on web scraping legality.

FAQ

How accurate is the higher education risk score compared to official accreditation status? The composite score is a public-data risk model, not an accreditation determination. It has not been calibrated against a ground-truth dataset of closures or sanctions. Think of it as an early warning indicator — it identifies institutions showing patterns historically associated with closure risk — not as a replacement for accreditor judgment. For institutions flagged ELEVATED or higher, the structured signals provide specific areas for human review.

How many institutions can I assess in one run? There is no hard limit per run. The practical limit is your per-run spending budget. At $0.045 per dossier call, $5 of Apify credits covers 111 institution assessments. For large portfolios, set a spending limit and batch institutions across multiple scheduled runs.

Does the higher education risk assessment cover community colleges? Yes. The tools work with any accredited institution type — research universities, liberal arts colleges, community colleges, technical schools, and for-profit institutions. Research integrity scores will be low for teaching-focused schools by design; interpret them relative to peer institutions of the same type using compare_peer_institutions.

How is this different from College Scorecard or IPEDS data? College Scorecard and IPEDS publish institutionally self-reported data with a 1-2 year lag. This server pulls live public signals from CFPB, Federal Register, and spending databases that reflect current activity. The combination captures complaints and regulatory actions that appear 12-18 months before they surface in official statistics.

Does this assess gainful employment metrics? The track_regulatory_actions tool monitors Federal Register publications for gainful employment rule changes and enforcement notices. Actual program-level debt-to-earnings ratios require Department of Education program-level data that is not available through the public APIs this server uses.

Can I use this server to compare for-profit and nonprofit institutions? Yes, and this is one of the most useful applications. The Accreditation Risk model specifically detects for-profit complaint and lawsuit patterns. Running compare_peer_institutions across a mix of institution types reveals whether a for-profit school's complaint and regulatory profile is typical of its sector or an outlier.

How long does a typical full dossier take? generate_institution_dossier fans out to 8 actors running with 256 MB memory and 120-second timeouts each, executing in parallel. Most dossiers complete in 60-120 seconds. Occasionally a single slow data source extends this to 3 minutes.

Is it legal to use this data for institutional risk assessment? All data sources are public government databases (CFPB, Federal Register, USAspending, Grants.gov) and open academic registries (OpenAlex, ORCID, Wikipedia). Accessing and analyzing this data is legal. When using outputs for consequential decisions affecting institutions, ensure your processes comply with applicable administrative and procedural requirements. See Apify's guide on web scraping legality.

Can I schedule this server to run portfolio sweeps automatically? Yes. Use Apify's built-in scheduling to run the MCP on a quarterly or annual cadence. Combine with a webhook to push results into a Slack channel, Google Sheet, or risk management system when the run completes. Contact Apify support for enterprise scheduling configurations covering portfolios of 500+ institutions.

What happens if one data source is unavailable during an assessment? The server uses Promise.allSettled for all parallel data collection. If a single data source fails or times out, the assessment continues with the remaining sources and returns empty arrays for the failed source. Scores will be lower than usual (a missing data source cannot contribute points), and the result will include the available signals from the sources that responded.

How is this different from hiring an education research analyst? A research analyst manually searching CFPB, Federal Register, USAspending, OpenAlex, and ORCID for one institution takes 4-6 hours at typical analyst billing rates of $50-150/hour. This server returns structured results in under 2 minutes at $0.045. For routine monitoring across large portfolios, automation reduces cost by 99% and eliminates the lag time between request and delivery.

Help us improve

If you encounter unexpected results or data quality 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, enterprise portfolio monitoring configurations, or accreditation agency licensing, reach out through the Apify platform.