MCP App Store
by KeenEthics
HeronHex-HigherEd icon

HeronHex-HigherEd MCP Server for Claude

By Heron & Hex Solutions

Overview

Ask about any U.S. college or university and get answers built only from free public sources, each one named and dated. Twenty-nine read-only tools cover five areas. Institutions: find a college by name or identifier, its profile, fall enrollment trends, retention, IPEDS finances, peers and side-by-side comparisons (IPEDS via the Urban Institute, NCES, College Scorecard). Risk: Heightened Cash Monitoring status, financial responsibility composite scores, cohort default rates (Federal Student Aid) and Single Audit findings with their text (Federal Audit Clearinghouse). Accreditation: who accredits an institution or program and the dated actions an agency has recorded (U.S. Department of Education DAPIP). Funding: research spending (NSF HERD), NSF and NIH awards, open Grants.gov opportunities, state higher-education funding (SHEEO) and projected high school graduates by state (WICHE). Labor market: wages by state or metropolitan area and national and state projections (BLS, Projections Central), occupation skills (O*NET), the program-to-occupation crosswalk, and College Scorecard program awards and earnings. What makes it trustworthy: every response carries its source, the period it covers and when it was retrieved; missing data comes back as an explicit null with the reason, never an estimate; identifier matches are exact and name matches are labeled as fuzzy and never guessed when ambiguous; upstream names and narratives are returned as data and flagged as untrusted. The tools report what the record says and do not score, rank or give verdicts. No sign-in and nothing to install. The server keeps no record of what anyone asks: no access log, no accounts, a cache of public data files only. Open source (MIT) from Heron & Hex Solutions; the same code runs locally as a Claude Code plugin if you prefer. Not an Anthropic product, provided without warranty, and not a substitute for the primary source on anything consequential.

How to connect

  1. Claude: Open the HeronHex-HigherEd listing in the Claude connectors directory and choose Connect. Claude listing HeronHex-HigherEd exposes 29 tools here.

Tools

compare_institutions

Claude
Side-by-side facts for 2 to 5 named institutions (by UNITID): control, state, Carnegie basic class, fall enrollment for the latest enrollment_years years, the latest IPEDS finance year (revenues, expenses, state appropriations, net tuition and fees), and College Scorecard cost and outcome fields. Institutions are compared as given; no peer matching, ranking or scoring is done. Years and accounting bases can differ between institutions, and the response says where they do. An institution that cannot be found is reported as a null and left out of the table.

compare_peers

Claude
Closest peers and their enrollment trends in one long table. Peer rule (deterministic, returned in the response): active institutions with the same control and 2025 Carnegie basic class (2021 class if 2025 is missing), ranked by closeness of latest fall enrollment (absolute log ratio). The target's own system is not excluded. scope narrows the pool: 'same_state' or 'same_system' (the target's system name in the IPEDS directory, for example a state community college system). Within a narrowed scope the Carnegie class is not required (the pool is already small), so peers match on control and institution level only; the returned rule says which was used. The default 'national' is size-matched anywhere in the country, which is not the same as a market or system benchmark.

find_program_offerings

Claude
Institutions that report awards in one program (4-digit CIP) at one credential level in College Scorecard field-of-study data, optionally within one state, ranked by awards (two consecutive award years added together; ties share a rank). Pass unitid to see where one institution sits in that list. Award counts include second majors and are not enrollments. Institution earnings are often suppressed for privacy; those are nulls with a reason. A program an institution offers but did not award in is not listed. When more than 1,500 institutions match, only part of the list can be read, so it is marked incomplete and no ranks are given.

get_accreditation_actions

Claude
Dated accreditation actions DAPIP holds for one institution by 6-digit IPEDS UNITID, newest first: the agency's own wording for each action (renewal, probation, show cause, heightened monitoring, loss of accreditation, program approval removed and so on), its stated justification, and a category that groups the wording. category is this tool's grouping, not the Department's or the agency's: loss, adverse, monitoring, resolved, routine, substantive_change, closure, agency_recognition, or unclassified for wording it has no rule for. Counts by category are returned for everything that matched before the limit applies. There is no score or verdict. A justification of "Other - Provide unlisted or multiple justifications" means the agency gave no listed reason; justification_detail is the agency's free text when present. Branch campuses are not included.

get_accreditation_status

Claude
Every accreditation record DAPIP holds for one institution by 6-digit IPEDS UNITID: the institutional accreditor, programmatic accreditors (for example nursing or radiologic technology) and the state agencies DAPIP carries (some state boards of nursing and career-technical boards). Each record has the agency, program, first accredited date, the status as DAPIP recorded it, the review date, the end date, current, and last_action (the newest dated action DAPIP holds for that record, or null). current is computed: DAPIP does not update the status text when a record ends, so a record with a passed end date and the status "Accredited" is not current. An ended record is not necessarily a loss of accreditation: an agency can be replaced by another, so read last_action and the actions list before saying why a record ended. action_history gives the earliest and latest dated action DAPIP holds for the institution, which bounds what the data can say about whether an accreditation was ever lost. scope is all, institutional, programmatic or state (records from state agencies). The institution's own record only: branch campuses and additional locations are not included. State board approvals outside the few state agencies in DAPIP are not in this data.

get_audit_findings

Claude
Single Audit (Uniform Guidance) submissions and findings for one auditee from the Federal Audit Clearinghouse. Give exactly one of ein (9 digits) or uei (12 characters); resolve_institution in highered-institutions returns both. Returns the latest years submissions with opinion, internal control and compliance flags, the filing timeline against the nine-month rule, and each finding's structured flags, decoded compliance-requirement labels and federal program. A finding's modified_opinion flag is FAC's, and FAC does not say which opinion it refers to; the financial statement opinion is gaap_opinion. Matches are exact on the identifier; there is no name search. A public college can report an EIN shared with a state or system, in which case an EIN lookup returns that parent's audit; a UEI lookup returns the college's own. Finding narratives are not returned here; get_finding_text returns them for one report, and report_url opens the full report. Some public institutions are audited inside a state or system-wide audit under a different identifier, so an empty result does not mean there was no audit.

get_enrollment_projection

Claude
Actual and projected HIGH SCHOOL GRADUATES by graduating class, for one state, from WICHE's Knocking at the College Door (11th edition, Dec 2024). This is a state-level supply indicator of future college-going students, not an enrollment forecast for any institution. Each row says whether it is reported, estimated or projected. Licensed CC BY-NC-SA 4.0 (non-commercial, share-alike): cite WICHE and respect the license when reusing.

get_enrollment_trend

Claude
Fall headcount enrollment (all students) by year for one institution, as a table. Default window is the latest 10 years. Years with no data are null with a reason.

get_financial_snapshot

Claude
Headline finances for one institution from IPEDS finance surveys, as reported, in dollars. Covers public institutions too (unlike FSA composite scores). Values are never scored or estimated. years > 1 adds a year-by-year history ending at the latest published year. One year's surplus or deficit is not a trend, so the history is the basis for any comparison over time. Institutions report on GASB-style (net position) or FASB-style (net assets) statements, whichever they use (public institutions may use either); the two are not directly comparable and the response names which one was reported. revenues_minus_expenses is computed here as reported total revenues minus reported total expenses; it is not an official surplus measure. Total revenues can include capital appropriations or gifts depending on the filer and year, so the reported component lines are returned and no 'excluding capital' figure is computed.

get_finding_text

Claude
The written narrative for findings in one Single Audit report, from the FAC findings_text table (what the auditor found: condition, criteria, cause, effect, recommendation), and the auditee's planned corrective action for the same finding, from the FAC corrective_action_plans table. Give the report_id from get_audit_findings, and optionally one reference_number such as 2024-003; otherwise the first limit findings are returned. Both texts are free text written by the auditor and the auditee. They can name staff, and this server cannot filter that. Each is cut at 6,000 characters (truncated says so), control and invisible formatting characters are removed, and charts or tables in the original are not included (contains_chart_or_table says so). Matches are exact on the report id; there is no search by text or name.

get_institution_profile

Claude
Profile of one institution: identifiers (UNITID, OPEID, EIN, UEI, domain), system and same-OPEID campuses, Carnegie codes, latest fall enrollment, College Scorecard outcomes. Missing data is null with a reason. identifiers.cbsa is the 5-digit metropolitan area code that highered-labor-market get_wages takes.

get_occupation_outlook

Claude
BLS Employment Projections 2025-35 for one occupation (SOC 2018): employment in 2025 and 2035, average annual openings (growth plus replacement), median wage, and typical entry education, experience and training. National figures only. Needs HIGHERED_MCPS_CONTACT (your email address), which BLS requires in the User-Agent and which is sent only to bls.gov. Some SOC codes are reported by BLS only as a broad group; the response says when that happened.

get_occupation_skills

Claude
Top-rated skills, core tasks and in-demand technologies for one occupation (SOC 2018) from the ONET database (CC BY 4.0, no key needed). Skill importance and level are ONET's 1-5 and 0-7 scales. Returns the base ONET occupation when one exists, and lists any specialties. Cite ONET as returned in citation.

get_program_alignment

Claude
Puts one program beside labor-market demand: the institution's awards and earnings for a 4-digit CIP and credential level (College Scorecard), the occupations the CIP-SOC crosswalk links to it, and for each occupation BLS national projected annual openings, typical entry education and median wage. It reports openings two ways, for all mapped occupations and for the subset whose typical entry education corresponds to the credential (a disclosed rule), and never reduces them to one score or ratio. Needs HIGHERED_MCPS_CONTACT for the BLS part; without it the rest is still returned. With state it also adds that state's own projections (Projections Central) for each mapped occupation beside the United States figures from the same source and years; these are a different period from BLS national 2025-35 and are never combined with it.

get_program_outcomes

Claude
An institution's programs from College Scorecard field-of-study data: 4-digit CIP, credential level, IPEDS award counts for two consecutive award years, and median earnings one year after completion. Institution earnings are often suppressed for privacy (small groups): those are nulls with a reason, and national medians for the same field are returned under their own label, never in place of the institution's. Sorted by awards, largest first.

get_research_spend

Claude
Total R&D expenditures by source of funds for one institution from the NSF HERD Survey, latest years survey years, in dollars (HERD publishes thousands; converted here). Join is on the exact UNITID. HERD covers institutions with at least $150,000 in R&D, so many small colleges are absent: that is a null, not zero. A few HERD institutions have no UNITID on file; they cannot be joined. Pass state to list those in that state (unjoined, for you to judge).

get_retention

Claude
Fall-to-fall retention for first-time degree-seeking undergraduates at one institution, from IPEDS Fall Enrollment part D, read straight from NCES's complete data files (no Urban API). For each fall collection: the full-time and part-time retention rate in percent as IPEDS publishes it, the number retained and the adjusted prior-fall cohort it is computed on, the student-to-faculty ratio, and IPEDS's flag for each value (reported, or how it was imputed or why it is missing; flag_meanings decodes them). The retention rate is the share of the prior fall's cohort (minus exclusions, plus inclusions since fall 2016) that is enrolled at the same institution the next fall, full- or part-time. It is not a graduation rate, and a student who transfers out counts as not retained. The rate is not recomputed here. The newest collection NCES has published can lag the present by a year or more.

get_state_funding

Claude
State and local support for public higher education by fiscal year, from SHEEO's State Higher Education Finance (SHEF) report: education appropriations, net tuition, total education revenue, student share, net FTE, state financial aid, and two-year vs four-year state and local support (FY2019 onward). State level only: this is not institution-level funding. Dollars are unadjusted; divide by an adjustment factor (HECA, COLI, EMI or CPI, returned per row) for real dollars, applying only one inflation adjustment.

get_state_occupation_outlook

Claude
A state's own employment projection for one occupation (SOC 2018) from Projections Central, the State Projections Consortium's public service: base and projected employment, change, average annual openings, and the projection years (typically 10 years for 'long', 2 years for 'short'). The United States row from the same source and years is returned for a like-for-like comparison. It covers a different period from BLS national 2025-35 numbers and is not comparable with them. State agencies produce these projections and are responsible for them. Some states report an occupation only as a broad group (the response says when). An occupation a state has not submitted is a null with a reason. Regional (sub-state) projections are not available here; the response links the state agency's own page.

get_title_iv_status

Claude
Title IV status for one institution by 8-digit OPEID: Heightened Cash Monitoring (HCM) list, financial responsibility composite score, and official cohort default rates. HCM and composite scores join on the exact 8-digit OPEID. Default rates are published by 6-digit OPEID, so they cover every campus sharing that prefix. Each section has its own source and as_of. reference quotes the federal default-rate sanction rule with its citation; it is the rule's text, not a finding about the institution.

get_wages

Claude
What workers in an occupation earn in a state, or in the metropolitan area around an institution, from BLS Occupational Employment and Wage Statistics (OEWS): for one occupation (SOC 2018 code; if you have only its name, find the code with search_occupations) in one state or one metropolitan area: employment, annual and hourly mean and median, annual 10th, 25th, 75th and 90th percentiles, and each estimate's relative standard error. Give exactly one of state (two letters) or area_code (the 5-digit CBSA code in an institution's cbsa, or a 7-digit nonmetropolitan code from search_metros). OEWS covers wage and salary workers in the occupation (not self-employed workers), of every experience level and credential, so it is not what recent graduates earn. Employment is rounded to 10. A cell BLS did not publish is a null with a reason; a wage at or above BLS's top code is returned as the top code with top_coded listing it, meaning the true figure is at least that. Needs HIGHERED_MCPS_CONTACT (your email address), which BLS requires in the User-Agent and which is sent only to bls.gov.

map_program_to_occupations

Claude
Occupations (SOC 2018) that NCES and BLS link to a program code (CIP 2020), 6-digit like 51.3801 or 4-digit like 51.38 (the union of its 6-digit codes). The crosswalk is many-to-many and not based on empirical data: it lists occupations a program may lead to, not where graduates actually work. Codes with no match return a null with a reason.

resolve_institution

Claude
Resolve a name, UNITID, OPEID, EIN, UEI, or web domain to ranked institution candidates. Returns each candidate with the basis for the match, including its metropolitan area code (cbsa, null outside any metropolitan area), which highered-labor-market get_wages takes for wages near the institution. If more than one institution matches, resolved is false and nothing is chosen. 6-digit numbers are checked as UNITID and OPEID6.

search_funding_opportunities

Claude
Open or forecasted federal funding OPPORTUNITIES from Grants.gov (not awards). eligibility filters to opportunities that Grants.gov codes as open to public or private institutions of higher education; a listing with no such code (several NSF solicitations, including some written for two-year colleges) is excluded by the filter, so a filtered list can be incomplete and any shows more. keyword is Grants.gov's own broad text match: a listing returned for a keyword does not necessarily concern it. Newest first. Opportunities are solicitations, not money an institution has received. A close date several years away is usually a continuing solicitation, not a deadline.

search_grant_awards

Claude
Federal research awards to one institution from NSF and NIH, identified by UEI. NSF: exact match on UEI; filter by award date (awarded_from/awarded_to). NIH: RePORTER cannot search by UEI, so it is queried by institution_name (NIH-style wording works best, e.g. 'PENNSYLVANIA STATE UNIVERSITY') for one fiscal_year and every row is kept only if its own UEI matches yours. NIH needs institution_name; without it that source returns a null with a reason. Awards only; use search_funding_opportunities for open solicitations. Principal investigators and program officers are never returned.

search_institutions

Claude
List institutions by filters (name text, state, control, 2025 Carnegie basic class code). Returns a table. Carnegie labels are not included; codes only.

search_metros

Claude
Metropolitan and nonmetropolitan areas in the BLS wage file whose name contains the text, with the code get_wages takes. Optionally limited to areas whose primary state is state. At most 25 are returned, in name order. An institution's own metropolitan code is its cbsa from resolve_institution; this tool is for finding a nearby area or a nonmetropolitan code.

search_occupations

Claude
Find the SOC code of an occupation from words in its title (for example registered nurses, welders, or dental hygienists). Returns each 6-digit SOC code and BLS's title, which is the code get_wages takes. A word matches its stem, so "nursing" finds "Registered Nurses". Only detailed occupation titles in the BLS wage file are searched, not job descriptions, so an occupation can sit under a different name than the one searched (nursing assistants, for example, are "Nursing Assistants" and "Orderlies").

search_programs

Claude
Find CIP 2020 program codes whose official title contains every word of the keyword (for example nursing, welding, or computer science). Returns each 6-digit code, the 4-digit code that College Scorecard program tools use (cip4), the title, and how many occupations the NCES crosswalk links to it (0 means NCES lists no match). Titles only are searched, not program definitions, so a field may sit under a different name than the one searched.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/heronhex-highered"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/heronhex-highered)