Candidcost is listed as active in the registry but did not answer our last check. It exposes 10 tools.
Cited, receipt-backed US home-buying data & calculators — education-only, every number sourced.
The linked repository no longer exists on GitHub — it was deleted or made private.
Uptime history
47 days of history · worst day 1%
47 days agonow
20.9%
Uptime 24h
19 of 91 checks
10
Tools
read from the server
386 ms
Response time
average over 24h
open, no key
Access
streamable-http
Candidcost does not always answer
Over the last week it answered 21.4% of our checks. We check every 15 minutes, so you hear about the next outage within the hour — not from your users.
Connect this server
Endpoint below is the one we actually reach during checks — not the one copied from a README. Last verified 10 min ago.
run in your terminal
claude mcp add mcp --transport http https://candidcost.com/api/mcp~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"mcp": {
"url": "https://candidcost.com/api/mcp"
}
}
}~/.codex/config.toml
[mcp_servers.mcp]
url = "https://candidcost.com/api/mcp".cursor/mcp.json
{
"mcpServers": {
"mcp": {
"url": "https://candidcost.com/api/mcp"
}
}
}.vscode/mcp.json
{
"mcpServers": {
"mcp": {
"url": "https://candidcost.com/api/mcp"
}
}
}Available tools 10
Read directly from the server with tools/list, grouped by what they act on.
If a tool disappears, we record the date.
affordability
affordability_estimate
Compute front-end (housing) and back-end (total debt) DTI budgets from income and debts, and the home price implied by the binding limit. Reports the standard 28/36 thresholds without ever telling the user what they "can afford". Dollar inputs are US dollars; money in the response is integer cents.
benchmark
benchmark_rates
Return the latest benchmark mortgage rates (e.g. 30-year and 15-year fixed) from the Federal Reserve (FRED), each with its observation date and ledger receipt. Rates are fixed-precision percent strings.
decode
decode_loan_estimate_fee
Classify a mortgage Loan Estimate fee by CFPB section and control class (set by lender, shoppable, third-party, government, or prepaid), with a plain-language explanation and an educational question a buyer could ask. Omit "fee" to list the full catalog.
dpa
search_dpa_programs
Find down payment assistance (DPA) programs by US state and/or buyer type. Includes national programs. Each result links to its official government source and carries a ledger receipt and last-verified date.
fee
fee_benchmarks
Return closing-cost and rate-spread percentile distributions (p25/p50/p75 + mean) for a US state, computed from the row-level CFPB HMDA LAR for originated, first-lien, home-purchase loans. Cells with fewer than 25 records are suppressed. Dollar metrics are integer cents (…Cents); percent metrics are fixed-precision strings (…Pct). Each distribution carries a ledger receipt hash for verification.
payment
payment_shock_estimate
Estimate why a mortgage payment often jumps in year two: escrow at closing is sized from the seller's current tax bill, but the county reassesses at your purchase price, and RESPA spreads the resulting escrow shortage over 12 months. Given the purchase price, the seller's current annual tax (from the listing), a state (or county FIPS) for the reassessment rate, and insurance, returns the year-1 escrow, the year-2 ongoing escrow, the temporary catch-up peak, and the jump. Money is integer cents. This decomposes the mechanics; it is not a prediction of your specific bill.
pressure
pressure_check
Decompose the tradeoff a rushed buyer faces: the earnest money at risk by walking away versus what overpaying to "win" costs over the life of the loan. Given the earnest money at risk, the overpay amount over your walk-away price, down-payment percent, rate, and term, returns the cash and financed shares of the overpay, the extra monthly P&I, the lifetime cost of overpaying, and its ratio to the earnest money. Money is integer cents. Educational decomposition — it names the two numbers so they can be compared; it never tells the user to walk or stay.
rent
rent_vs_buy_estimate
Compare the net cost of owning against renting over a holding period and find the break-even year. Owning nets upfront cash and recurring carrying costs against sale proceeds (appreciated value − selling costs − remaining loan); renting nets rent paid against the investment gain on the cash a buyer would tie up. Dollar inputs are US dollars; money in the response is integer cents. Appreciation, rent growth, and investment return are user-set assumptions. Tax deductions are excluded (documented simplification). Educational decomposition — it reports both net costs and the break-even year; it never says rent or buy.
true
true_cost_estimate
Estimate the full monthly cost of owning a home — principal & interest, property tax, insurance, PMI, HOA, and maintenance — and contrast it with the lender "approval" payment (PITI+PMI) to reveal the hidden monthly gap. Dollar inputs are US dollars; all money in the response is integer cents; rates are fixed-precision percent strings.
verify
verify_receipt
Verify a CandidCost ledger receipt hash (full sha256 or short "3f9a…c21e" form). Confirms the receipt exists and that its dataset hash chain is intact, proving the underlying data has not been altered since ingestion.
Endpoints
| URL | Transport | State | Latency | Checked |
|---|---|---|---|---|
| https://candidcost.com/api/mcp | streamable-http | answering | 37 ms | 10 min ago |
Alternatives to Candidcost
same job, measured the same wayVerified Numbers
by jfortier93
Current source-cited 2026 US tax constants and calculators for AI agents; every answer cites IRS.
eDiscovery Decoder News/Calc
by ediscoverydecoder
Free educational MCP for read-only eDiscovery news and TAR/review calculators.
Proofetch
by vektoris-ai
Verified extraction: source-backed JSON from PDFs/URLs; honest null + signed receipt.
Hive Intelligence MCP
by hiveintelligence
Evidence-backed crypto due diligence with sources, freshness, and a runtime receipt on every call.
Greencalculus
by greencalculus
Sourced carbon emission factors + audit-traced calculations an AI can cite.
MyPlate.food nutrition tools
by myplate
USDA nutrition calculators + the preserved MyPlate Kitchen recipes; source-cited, machine-verified.
Agentcite
by kevinx702
Fabrication-free, DOI-backed citations for AI agents — real openAlex sources, never hallucinated.
AI Vendor Data Trust Index
by oscarlopez-gif
Source-cited data-handling comparison of 11 AI vendors: training, retention, opt-out.
Candidcost — questions
Answers built from our own checks of this server.
What can Candidcost do?
It exposes 10 tools, read directly from the server on our last check. Among them: affordability_estimate, benchmark_rates, decode_loan_estimate_fee, fee_benchmarks, payment_shock_estimate, pressure_check and 4 more. The full list with descriptions is on this page — we take it from the server itself via tools/list, not from a README.
How MCP servers expose tools in the first place →
Is Candidcost working right now?
We send a real MCP handshake every 15 minutes. Over the last 24 hours 19 of 91 checks got a reply (20.9%), average response time 386 ms. The bar chart above shows every period we have measured.
Is Candidcost still maintained?
The linked repository no longer exists on GitHub — it was deleted or made private. We show this because it changes what you can expect: an unmaintained server may keep answering for months and then stop without warning.
The registry lists Candidcost as active — why does it not respond?
The official MCP registry stores what the author submitted; it does not verify that the server still runs. We check the endpoint ourselves, and this one does not answer. Catalogues that copy the registry without checking will show it as working.
How do I connect Candidcost?
Copy the ready config from this page — we generate it for Claude Code, Claude Desktop, Codex, Cursor and VS Code, each with the file path that client actually reads. It is a remote server, so there is nothing to install — the client connects to the address.
Does Candidcost need an API key?
No. Candidcost completed a full MCP handshake with us as an anonymous client and listed its tools without asking for anything. All 10 of them are readable on this page. This is what we observed, not what the docs claim.
How fast is Candidcost?
It answers our handshake in 386 ms on average, which is faster than 39% of all working MCP servers we measure. The comparison comes from our own checks across the whole registry, every 15 minutes.