About college-roi-mcp
college-roi-mcp is a read-only MCP server for exploring U.S. college and major return-on-investment data. It provides college rankings and ROI records without an API key, using data from a free API.
Key features
- 01Search 500 four-year colleges
- 02Compare 30-year college NPV
- 03Review lifetime ROI by major
- 04Rank state best-value colleges
- 05Compare out-of-state tuition penalties
Use cases
- 01Compare ROI across majors
- 02Find a state's best-value colleges
- 03Estimate the long-term cost of nonresident tuition
MCP support
The server exposes read-only tools for searching colleges, majors, rankings, and API metadata. MCP clients can run it over stdio; the underlying data API requires no key.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
7Saved tool list · updated 2w ago
This tool takes no parameters.
Details
From the project READMEcollege-roi-mcp
MCP server for U.S. college-ROI data: 30-year net present value (NPV) for 500 four-year colleges, lifetime ROI for 19 major categories + 115 CIP subfields, per-state best-value rankings, and the out-of-state tuition penalty ranking.
Wraps the free, keyless LE TEEN College ROI API (static JSON, CC BY 4.0). No API key, no signup, no rate limits beyond CDN sanity. Data derives from FREOPP program-level ROI estimates, IPEDS institutional data, and BEA price deflators — methodology at le-teen.com/methodology. Dataset DOI: 10.5281/zenodo.21351602.
Install
Requires Node ≥ 18.
Claude Desktop / Claude Code (claude_desktop_config.json or .mcp.json):
{
"mcpServers": {
"college-roi": {
"command": "npx",
"args": ["-y", "college-roi-mcp"]
}
}
}
Claude Code CLI:
claude mcp add college-roi -- npx -y college-roi-mcp
Any other MCP client: run npx -y college-roi-mcp over stdio.
Tools
| Tool | What it returns |
|---|---|
list_majors | Compact ROI rows for 19 categories / 115 subfields — median & mean lifetime ROI, share of graduates in programs that never break even, median break-even age among programs that break even, AI-exposure band (category rows only). Filter by kind/parent, sort, limit. |
get_major | Full record for one major: ROI distribution (p25/median/mean/p75), completion-adjusted & dropout ROI, graduates/programs, AI-exposure detail. |
search_colleges | Search the 500 positive-NPV-envelope colleges by name/state/control — resident & non-resident 30-yr NPV, cost of attendance, median earnings, break-even age. |
get_college | Full record for one college, incl. IPEDS unitid and FREOPP coverage. |
best_value_colleges | Each state's best-value four-years at resident pricing (top 15 per state), or a nationwide top-pick summary. |
out_of_state_penalty | Public four-years ranked by how much 30-yr NPV a non-resident gives up vs a resident. |
get_api_info | Metadata for the underlying API: endpoints, license, DOI, attribution. |
All dollar figures are lifetime/30-year totals in USD. List responses are compact projections with limit controls so they stay small in agent contexts; detail tools return the full upstream record.
Note the envelope: this server's college list holds only the 500 four-year colleges with the largest FREOPP program coverage that also clear our data gates and show a positive 30-year net value. Absence from the list is a coverage statement, not a verdict.
Example prompts
- "Is a nursing degree worth it compared to computer science?"
- "Best-value colleges in Michigan for a resident."
- "How much does going out-of-state to UVA cost me over 30 years?"
- "Which majors have the worst odds of ever breaking even?"
Data freshness & caching
The upstream API is static JSON regenerated with the dataset; this server caches responses in-process for 24 h (matching the API's CDN cache). Nothing is written anywhere; the server is read-only.
License & attribution
- Code: MIT.
- Data: CC BY 4.0 — free to use with attribution. Cite LE TEEN, le-teen.com (and FREOPP/IPEDS/BEA as upstream sources).
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.