One deal file — address to close
One deal file. Everyone sees the same work.
Address → intelligence → sizing → stress → memo → lenders → term sheets → close
- Brokers —
- stop guessing which lenders bite. See who clears the deal before you send it.
- Lenders —
- only deals that clear your criteria reach you. Your box decides — test a change against your last ten deals before it takes effect.
- Borrowers —
- one application, one document set, and a timeline of what each lender is doing. No black box, no sending the same file five times.
Free for borrowers. Start with an address and the basics.
Free to run deals. Pay only when one closes.
An address is the whole intake. Everything below is derived, not typed.
The first question a lender asks
AI in the loop, never in the chair.
It reads the documents and drafts the memo. Your policy and a deterministic engine decide. Every number traces back to the page it came from.
Every figure links to the section that demonstrates it.
One deal. Start to finish.
Below is a single sample deal — a 24-unit mixed-use refinance in Brooklyn — run through the same engine that runs live deals. The numbers are computed, not invented. Check the math; it reconciles.
24-unit mixed-use · Refinance
Ask: $4,850,000 · Est. value: $7,460,000
Fictional property, for demonstration. Same engine, same outputs — your numbers when you run it.
A deal file in 90 seconds, not 90 emails.
Type an address. Relendi pulls valuations, flood zone, demographics, violations, environmental flags, and rate context from 12+ government and market data feeds. That's day one of package assembly — the data nobody should be copy-pasting — done before the kickoff call.
Start your deal — bring an addressCount them — 12. NYC Open Data and the FL county appraisers join on their own deals; this one is in Brooklyn. Named as sources, not partners — none of them endorse Relendi.
| Scenario | Rate | NOI | DSCR | Status |
|---|---|---|---|---|
| Base case | 6.85% | $560,000 | 1.47x | — |
| Rate +200bps | 8.85% | $560,000 | 1.21x | PASS |
| Rate +300bps | 9.85% | $560,000 | 1.11x | MARGINAL |
| Income -15% | 6.85% | $476,000 | 1.25x | PASS |
| Income -25% | 6.85% | $420,000 | 1.10x | MARGINAL |
| Vacancy +15pts | 6.85% | $476,000 | 1.25x | PASS |
| Combined Stress | 8.85% | $476,000 | 1.03x | FAIL |
| Severe Stress | 9.85% | $420,000 | 0.83x | FAIL |
We show the failures. So should your deal package.
Vacancy +15pts rebuilds NOI through the same underwriting path — a different shock from Income −15%, not a relabel of it. Identical scenarios are collapsed, so nothing counts twice, which is why the number of rows depends on the policy rather than on us.
Stress-tested before a lender ever asks.
Every deal gets an AI risk analysis and a stress battery the lender's own policy defines — rate shock, income decline, vacancy and combined downside — so the first hard question a lender asks is one you've already answered. Including the scenarios that fail. Especially those.
Clears at $5,222,000
— bound by LTV. Against the $4,850,000 ask, that is $372,000 of headroom — not cash, and not an offer.
- DSCR ceiling · 1.25x$5,697,491
- LTV ceiling · 70%$5,222,000binds
- Debt-yield ceiling · 9.5%$5,894,737
Sizing takes the lowest of the ceilings the box implies, not the highest. This is the lender's view of that file — the exact figure and every ceiling behind it. The sponsor is shown the same answer and the test that set it, in the same words, rounded to the nearest $250,000: enough to act on, not enough to reverse-engineer the floors in someone's credit box.
Matched lenders. Compared on 5-year cost, not teaser rate.
The matching engine screens your deal against each lender's actual box — asset class, geography, leverage, DSCR floors — then compares competing term sheets on total 5-year cost of capital: rate, points, fees, amortization, prepay. The cheapest rate is not always the cheapest money. Now you can prove it to your client.
Run your own numbers.
Commercial lenders size a loan by taking the lowest of four constraints, not the highest: loan-to-value, debt service coverage, debt yield, and — on acquisitions and construction — loan-to-cost. Whichever binds first is your proceeds. Our underwriting engine sizes live deals on the first three and names the one that binds; this calculator adds loan-to-cost. Change any input and watch which constraint takes over.
Binding constraint: Loan-to-Value (LTV)
Your credit box, encoded — and testable before you save it.
Not three fields in a settings tab. Your targets, your decline floors, your per-asset-class overrides — the four tests a deal is actually run through, in the order the engine runs them, with both numbers on every rung: the one that flags a deal, and the one that declines it.
Change a number and your own last ten deals move on the ladder before anything takes effect. Version history records who changed what, rollback is one click, and only owners and admins can move the box.
See the dry run, the custody trail and the data promiseTen deals, resting on the rung that caught each.
Can the rent cover the payment?
DSCRflags below 1.25x · declines below 1.15x
- 240 Halsey St, Brooklyn, NY — declined
- 1815 Coral Way, Miami, FL — flagged for conditions
Are we lending too much against the value?
LTVflags above 70% · declines above 75%
- 700 NE 2nd Ave, Fort Lauderdale, FL — declined
- 88 Kent Ave, Brooklyn, NY — could not be tested
Is the yield on our money enough?
Debt yieldflags below 9.5% · declines below 8%
- 2201 NW 7th St, Miami, FL — declined
Does it survive a bad year?
Stressdeclines above 2 scenarios under 1.00x
- 740 Calverton Row (RX-2231), Brooklyn, NY — flagged for conditions
- 3400 Hillsboro Blvd, Deerfield Beach, FL — declined
Through every test
nothing caught them- 1470 Flatbush Ave, Brooklyn, NY — cleared every test
- 6120 Biscayne Blvd, Miami, FL — cleared every test
- 415 Ocean Dr, Miami Beach, FL — cleared every test
- Declined
- Conditions
- Untested — a file is missing
- Cleared
The file you have been watching is on rung 04. RX-2231 clears every threshold in this box, and one severe scenario dips under 1.00x coverage — so it is flagged for conditions, not declined. Keeping those two apart is the whole reason to encode a box instead of a cutoff.
And then it does a second job
Your box isn't only governance. It's the filter.
A deal reaches you because it cleared the policy you configured — screened on geography, asset class, leverage and the documents your policy requires, before anyone on your team opens the file. The alternative is the inbox: fifty a week, most of them wrong on something you had already written down.
We are not promising you volume. We are early, and we would have to invent the number. What the box lets us tell you is the other half — what stops at the door, and which of your own rules stopped it.
A committee-ready credit memo in 30 seconds.
Ten sections. Risk-forward. Built on the box you just configured and the assumptions behind it — vacancy floors, management fees, replacement reserves — not the broker's. Your analyst gets a first draft in about 30 seconds instead of a blank page and a lost day, and your file shows the documented independent underwriting your examiners ask for.
- ✓Executive Summary
- ✓Borrower/Guarantor Analysis
- ✓Property Analysis
- ✓Financial Analysis
- ✓Stress Testing
- ✓Risk Assessment
- ✓Collateral Analysis
- ✓Policy Compliance
- ✓Proposed Terms
- ✓Recommendation
| Line item | Broker | Lender defaults | Δ |
|---|---|---|---|
| Vacancy | 3.0% | 7.0% floor | −$35,200 |
| Management | 3.0% | 5.0% | −$17,600 |
| Reserves | $0 | $300/unit | −$7,200 |
| NOI | $560,000 | $500,000 | −$60,000 |
| DSCR | 1.47x | 1.31x | — |
The sizing moves with it, and so does the binding test. On the broker's $560,000 the file cleared at $5,222,000 with LTV binding. On your $500,000 it clears at $5,087,045 and coverage takes over.
A first draft for your credit team — not a credit decision. Every number traces to its source.
Signed term sheets aren't closed loans.
Roughly 30% of deals that reach a term sheet still die in closing coordination — appraisals nobody ordered, title items nobody chased, signature pages lost in inboxes. Relendi tracks all 23 closing milestones with owners across broker, lender, borrower, attorney, and title — so the deal that got this far actually funds.
Currently in progress: Survey Received — assigned to lender, due Friday.
Built for everyone at the table.
FOR BROKERS
A bigger book. No new hires.
- Know which lenders clear a deal before you send it — matched against their real credit box, not a profile they filled in once
- Files assemble from an address in under 90 seconds, intelligence attached — packages that don't bounce for missing data
- Threaded lender email, term sheets compared on 5-year cost, not teaser rate
Free to run deals. Success fee only when a lender we introduced funds one.
FOR LENDERS
Your credit box, encoded — and testable.
- Set your targets, decline floors and per-asset-class overrides, then test any change against your own last ten deals before it takes effect
- Deal flow that already cleared your policy — complete files, screened before a human opens them
- Version history, one-click rollback, and only owners and admins can move the box
We'll scope it to your institution.
FOR BORROWERS & SPONSORS
Find out what actually clears.
- A dollar figure and the name of the test that held the loan down — not a letter grade
- The bridge from your filed T-12 to the underwritten line, every deduction labelled
- One guided application, documents in one place, ESIGN signatures, anonymised lender activity on your deal
Free for borrowers.
What you won't find on this page.
No borrowed logos. No “trusted by” wall we haven't earned. No billions-originated counter. We're an early platform processing real deals in New York and Florida, and our proof is the machine you just watched run. When we publish a number, you can check it.
No model accuracy percentage — we would have to grade our own homework. No lead volume, because we would have to invent it. No odds of approval, because a lender decides that and we don't. And your credit box stays yours: not pooled into a benchmark, not shown to another lender, not published. What we do measure across firms is marketplace behaviour — how often a lender responds, issues a term sheet, or funds one — never the box itself.
Fair questions.
The deal is a sample; the math is real. Every figure — the 1.47x DSCR, the 0.83x severe-stress scenario, the NOI deltas — is computed by the same formulas that run live deals, and it reconciles if you check it. We'd rather show you one honest sample than a wall of logos we haven't earned.
No — it's generated b-roll, and it's the only thing on this page that is. Four ordinary buildings, no landmarks, nothing claiming to be a specific address. Everything with a number attached — the sizing, the stress scenarios, the memo — is computed by the same engine that runs live deals. We're telling you because you'd work it out, and because a page that says it shows its work should start with its own.
No — the initial sizing is indicative: preliminary numbers, rate context, and the questions a lender will probe, based on what you tell us. A real term sheet comes from a real lender after underwriting. We say this plainly because pretending otherwise is how this industry lost everyone's trust.
It's a first draft, not a credit decision. The memo is built on your underwriting assumptions — vacancy floors, management fees, replacement reserves — shown side-by-side against the broker's NOI with line-item deltas. Your analyst reviews, adjusts, and owns the conclusion. What your file gains is speed (a 10-section draft in ~30 seconds vs 6–10 hours) and the documented independent underwriting examiners look for.
Because we'd have to make them up, and you'd eventually find out. Relendi is an early-stage platform processing real deals in New York and Florida. Our credibility strategy is simpler: run the product in front of you and publish only numbers you can verify — 90-second intake, four tests on every deal, 23 closing milestones, 30-second memo drafts. Where a number is set by the lender rather than by us — how many stress scenarios a deal gets, for one — we say so instead of picking a figure that reads well.
Brokers: free to run deals — every step needed to close one is free, with no card. A success fee of 15 basis points applies only when a lender Relendi introduced funds your deal, floored at $2,500 and capped at $25,000. Close with a lender from your own book and it's $0. Broker Pro adds unlimited deals, unlimited AI and white-label at $149/month intro. Lenders: open a lender workspace and we'll scope your pipeline together. Borrowers: free.
New York and Florida today, where our property intelligence is deepest — ATTOM valuations, FEMA flood zones, Census demographics, NYC DOB violations, EPA environmental records, Florida county appraiser data, and more. Deals elsewhere can still run through the platform; the automated intelligence layer is strongest in NY and FL. We'd rather be excellent in two states than shallow in fifty.
Documents live in encrypted storage with pre-signed, expiring access links — never public buckets. E-signatures are ESIGN-compliant. The platform is built around ECOA, GLBA, BSA/AML, HMDA, and CAN-SPAM requirements, and every action on a deal is written to an audit trail.
Relendi inside your AI.
Add Relendi as a connector in Claude or ChatGPT and ask for a loan size, a stress test, or NY/FL closing costs in plain English. The math is deterministic — the assistant reports the numbers, it doesn't invent them. Free, no account, no API key.
https://relendi.com/api/mcp/publicYour deal, lender-ready. Start free.
Bring an address and the basics. Relendi assembles the deal file, runs the intelligence, and gets you in front of matched lenders — sizing, rate context, and what a lender will probe, shown with the work. That's the point.
Free for borrowers. Free for brokers to run deals. Already on Relendi? Sign in