LoanFactory, Inc

By LoanFactory, Inc

Overview

Get real mortgage numbers from Loan Factory, Inc. (NMLS #320841), a licensed mortgage broker, priced off today's lender rate sheets. Ask what a loan would actually cost you and get current options with the note rate, APR, monthly principal and interest, mortgage insurance where it applies, and estimated closing costs. Ask what you can afford on your income and existing debts. Check whether a loan is conforming or jumbo in a specific county, what the FHA ceiling is, and how the limits change for a duplex, triplex or fourplex. You do not need to know your ZIP code — give a city and state and it resolves the county for you, since loan limits are set per county and can differ by hundreds of thousands of dollars within one state. Answer with whatever you know. If you leave something out, a representative figure is used and reported back, so you always see what the numbers rest on and can correct it. Rates are estimates, not loan offers or locked rates, and no account or sign-up is required.

Tools

estimate_affordability

ChatGPT

Estimate the highest purchase price a borrower can carry, from their income, existing debts, cash for a down payment and a debt-to-income ceiling. Use this for 'how much house can I afford?', 'what can I buy on $X a year?' or 'is a $700k house realistic for me?'. Every argument is optional and a call with no arguments is valid: pass what the user said and a representative figure is used for the rest, then reported in assumptions. An income and a debt figure are enough to call with. This returns a computed maximum against a stated DTI ceiling, which is a firmer answer than an estimated range and shows its own workings. When assumptions is non-empty you must say what was assumed alongside the number, and offer to redo it with the user's real figures. Income and existing debt move this result more than anything else. For the interest rate, call get_mortgage_rate_estimate first and pass a real quoted figure — the default here is a placeholder that goes stale. This is arithmetic against a DTI ceiling, not an underwriting decision. It does not check credit, assets, employment history or program eligibility, and it does not model mortgage insurance.

get_county_loan_limits

ChatGPT

Look up conforming and FHA loan limits, plus area median income, for the county a US location sits in. Use this for any question about whether a loan is conforming or jumbo, what the FHA ceiling is, how limits change for a duplex, triplex or fourplex, or whether someone qualifies for income-restricted programs like HomeReady and Home Possible. Every argument is optional and a call with no arguments is valid: pass a ZIP code if one was given, otherwise a city and state, otherwise nothing — a representative county is used and named in assumptions. When that list is non-empty, tell the user which county the figures describe, because limits differ by hundreds of thousands of dollars between counties in the same state. Limits are set annually.

get_mortgage_rate_estimate

ChatGPT

Return current mortgage rate estimates for a US property, priced by Loan Factory's pricing engine against today's lender rate sheets. Every argument is optional and a call with no arguments is valid. The tool prices a representative scenario for anything not supplied and reports each substituted figure back, so a mortgage question can be answered as asked rather than turned into a list of follow-up questions. 'What are mortgage rates right now?' is a call with no arguments; 'what are rates in Austin?' is a call with a city. A price without a loan amount is enough: pass property_value alone and the loan is derived at 80% of it, with the derivation named in assumptions. A loan amount without a price works the same way. Neither needs to be asked for first — call with what was said, show what was assumed, and let the user correct it. Pass values the user gave and omit the ones they did not. Do not invent an argument to fill a gap — omitting it is how the tool knows to treat that value as assumed and report it back. The result returns an assumptions list naming every value that was defaulted and the figure used. When it is non-empty you must tell the user what was assumed, in the same reply as the rates, and offer to reprice with their real numbers. Credit score, ZIP code and down payment move these figures materially, so a rate presented without its assumptions is misleading. Give city and state instead of a ZIP code if that is all the user knows — the tool resolves it. Loan limits are set per county. The result includes details_url, which opens Loan Factory's rate tool on this exact scenario. Offer it when the user wants the full breakdown, wants to change an input, or wants more programs than are returned here. Always show the returned disclosure alongside any rate. These are estimates, not loan offers.

resolve_zip_code

ChatGPT

Turn a US city and state into a confirmed ZIP code and the county it sits in. get_mortgage_rate_estimate and get_county_loan_limits accept a city and state directly, so this is only needed when the ZIP code itself is the question, or when a city spans several counties and the user must choose between them. Do not guess a ZIP code from a state on your own. Conforming loan limits and area median income are set per county, and a single state contains many counties with very different limits — in California they range from 832,750 to 1,249,125. Picking the wrong one silently misprices the loan. Returned ZIP codes are checked against the pricing engine, so each one works and each county name is the engine's own. When a city spans several counties the result is marked ambiguous and lists each one: ask the user which county before quoting, and do not choose for them. A county Loan Factory does not lend in is reported with is_serviced false. Say so plainly instead of quoting a rate for it.