Every OM is the seller's case. Get the cross-examination in minutes.
Drop or forward an offering memo. DeepTenant grades it, checks the numbers against the T-12, the county tax rate, and today's debt market, and writes the request list for the broker.
No card. Unlimited Rapid Screens on every plan.
Illustrative synthetic deal: Maple Grove, 96 units, Papillion NE.
Send it the way you already get it.
Forward broker emails to your firm's address or drop the PDF. The OM, rent roll, and T-12 are read, filed, and screened before you open them.
A verdict you can defend to a lender.
Pursue, pass, or pursue below ask, with every number labeled Sourced, Computed, or Assumption, and an Excel model with live formulas your lender can change.
Measured against your own deals.
Log what you won, lost, and passed on. Each new screen shows where it sits against your history, which is the one comp set nobody can sell you.
Your assistant reads the deal. DeepTenant keeps the record.
Add DeepTenant as a connector in Claude or Cowork. Ask it to screen an OM, save a document, note what the broker told you, or log how a deal ended. The math runs on our engine and everything lands in your workspace, so the work you do in chat is still there next quarter.
- In Claude: Customize → Connectors → Add custom connector
- Paste
deeptenant.com/api/mcp - Sign in to DeepTenant and approve
On every plan, Free included. Also works with ChatGPT and any client that speaks MCP. Setup and tools →
Twelve checks that run on a cold OM, in any market.
- Pro forma presented as actual
- Rent roll × 12 against T-12 collections
- Taxes reassessed at your price
- Debt re-run at today's rate
- Expense ratio against market
- Unexplained add-backs
- Missing T-12 or rent roll
- Stale or distant comps
- Exit cap inside the going-in
- Understated insurance
- Conflicts between documents
- Material fields verified before the math
If a number can't be traced, it doesn't get printed.
The math is code, not a language model. Every figure carries one of four labels, and a missing document becomes a request, never a guess.
- Your data stays yours. Isolated per firm in the database and tested on every build.
- Personal data is dropped at the door. Tenant names and SSNs never reach the searchable layer.
- Every action is on the record. Who approved a number, what the AI changed, when.
Monthly, cancel any time, and no charge per deal.
Rapid Screens are unlimited on every plan, Free included. Plans differ only in seats and Full Evaluations, the deeper underwriting with the cited memo and the lender model.
Free
- Unlimited Rapid Screens
- 2 Full Evaluations a month
- Deal inbox, imports, Excel and Word exports
- Your deal history and the Claude connector
Pro
- Everything in Free
- 20 Full Evaluations a month
- For one person working deals every week
Team
- Everything in Free
- 75 Full Evaluations a month, shared
- One deal history the whole firm builds
- Each person connects their own Claude
Pro and Team open for subscription after the Founding Beta. Until then, start on Free or apply for Team free.
Use it free while you help shape it.
We're bringing on a small group of CRE pros who look at real deals every week. You get the product free and a direct line to the founder; we get to hear where it's wrong.
You get
- Team plan free for 3 months, up to 5 people
- 50% off for 12 months after that, if you stay
- Your deals and history kept either way
We ask
- Run at least three real OMs through it
- One 20-minute call around day 30
- Tell us where it was wrong
The short answers.
What's the difference between a Rapid Screen and a Full Evaluation?
A Rapid Screen grades the OM, runs the twelve checks, gives a verdict, and writes the broker request, in about a minute and free on every plan. A Full Evaluation adds the cited memo: pricing against comps, the market read, build-out costs, and the lender model. Plans include a monthly number of Full Evaluations; there are no per-deal charges.
Which markets does it know?
The checks run anywhere, on the documents plus national sources like the FRED rate. Cited local comps and benchmarks are deepest today for Omaha, Kansas City, and Wichita; elsewhere the screen says what it could not source instead of filling the gap.
Does it make numbers up?
No. Figures come from your documents, a cited source, or the calculator, and each one is labeled. An assumption you change is marked as yours.
Do I need Claude to use it?
No. The web app does everything on its own. The connector is for people who already work in Claude, Cowork, or ChatGPT and want their deals to land in one place.
Can I hand this to a lender or my partners?
Yes: a Word report and an Excel model with live formulas. It's decision support, not an appraisal or investment advice, and the report says so.