DealLens Brief: a one-page investor memo, generated from a single MLS link.
DealLens already does the AI valuation and comp analysis. The Brief is what an investor actually shares with a partner, an LP, or themselves on a Sunday night. One PDF, one screen, all the math, and a clear recommendation. Generated automatically from a single MLS or Zillow link in about 60 seconds.
The Challenge
The DealLens product turned a great underwriting workflow into a real one, but investors still had to copy the numbers into a Google Doc, paste in a property photo, write the thesis themselves, and send it to a partner. That re-typing step was where deals died: by the time the memo was written, the property had three other offers.
We needed an output artifact that worked for the way investors actually shared deals: PDF in iMessage, link in email, embed in a notion page. Branded, concise, and convincing enough that the partner could approve in 90 seconds without opening the listing.
What I Built
A generator that turns an MLS or Zillow link into a one-page investment brief.
Header: address, property type, schools, days on market.
Hero: a clean property silhouette (not a stock real-estate photo), key info, market chip.
Key row: list price, cap rate, cash-on-cash, 5-year IRR, all computed from the buyer's underwriting profile.
AI valuation panel: estimate, range, comps used, training base, and a confidence score.
AI thesis: four short bullet points written by the model, citing the comps and submarket data.
Comps strip: three closest comps with delta vs subject.
Recommendation stamp: BUY at $X, HOLD, or PASS with a one-line rationale.
The brief renders as a live web page (shareable URL), exports to a print-perfect PDF, and can be auto-emailed to a deal list.
Technical Foundation
Next.js for the live brief view and the public share URL.
FastAPI service that scrapes the MLS or Zillow listing, normalizes it, and runs the underwriting math.
Claude for the AI thesis bullets, prompted with the comp data and market stats so it cannot freelance.
A custom React-PDF renderer that takes the same JSX and outputs print-ready PDF with the right typography and the BUY stamp.
pgvector for "find me three closest comps that match buyer's box" plus a 412k Atlanta sales training set.
Resend for the email delivery, with branded templates and a one-click unsubscribe.
Outcome
Average time from listing link to shared brief: 58 seconds.
64% of generated briefs are shared at least once (iMessage, email, link).
Partner-approval cycle (time from brief sent to "yes/no") down from a stated average of 1.6 days to 4 hours.
Adopted by two real estate funds and a 14-agent investor sales team in the Atlanta market.
Like this project
Posted May 8, 2026
A one-page investor memo generated from a single MLS link. Comps, AI valuation, returns, and a buy/hold/pass recommendation in 60 seconds.