Help us direct you to the right place to sign up
Comp sourcing slows live deals.
Broker calls take days the process doesn’t have.
For Private Equity · CompStak One Connect
Offering memos are written to sell. CompStak shows what comparable leases actually netted and the cap rates behind recent sales, giving your model numbers that hold up at investment committee.
Seller’s offering memo vs. verified CompStak comps · Class A office, Gotham CBD
Trusted by the leaders in commercial real estate
Broker calls take days the process doesn’t have.
Every assumption inherits the seller’s framing.
A published average is not an edge.
CompStak One Connect answers each of these with verified, analyst-reviewed comps.
Get StartedUnderwrite on lease, sale, loan, and property data including NER, concessions, sale price, cap rate, NOI, and LTV. Portfolios keeps every position measured against the market through the hold.
Get StartedUnderwriting · 400 Mandel Street
Office, Class A · 1,240,000 SF · Gotham CBD · sold Apr 2025
Describe a space in Rent Predictor, including one with no lease on file, and get the Estimated Starting Rent Today with the comps behind it.
Get StartedRent Predictor · model input
400 Mandel Street · Floor 28 · 42,000 SF · no lease on file
Query CompStak comps from ChatGPT, Claude, Gemini, or Copilot through the Agent Connect MCP server, and join them against your own models and internal records. Statistics and trends are free; pulling the underlying records counts toward your weekly comp limit.
Get Started“What have Class A office buildings in Gotham CBD traded at over the last four quarters, and where do cap rates sit?”
9 sales closed since Q3 2025 at a median $679.03 per square foot. Median cap rate 5.10%, up 35 basis points year over year.
Rent Predictor and Agent Connect are available on CompStak One Connect.
Agent Connect puts CompStak inside Claude, ChatGPT, Gemini, and Copilot. Ask the way you’d ask an analyst, and the work comes back with its source comps attached.
“Build me an investment memo for 555 California Street: rent roll context, comp set, submarket trends, and recent sale comps.”
“Underwrite market rent for 55 Hudson Yards: median starting rent, free rent, and landlord work on comparable leases signed in the last 18 months.”
“Stress test my rent assumption of $85.23/SF for 600 Montgomery Street against leases signed in its comp set this year.”
“Pull sale comps for Class A office in Dallas since 2024 and summarize cap rates, price per SF, and who is buying.”
Access verified lease, sale, loan, and property data across major CRE asset types in 105 markets, with the details to analyze markets, value assets, and underwrite opportunities.
Across the asset types you work with:
*Multifamily data powered by RealPage
Each comp goes through a multi-step data verification process.
40,000+ verified CRE professionals at brokerage and appraisal firms submit comps daily. Transaction data is received and reviewed multiple times.
Machine learning and statistical anomaly detection flag figures outside the expected range for the market and building. CRE Ontology resolves the address to one building and matches every party on the deal to a verified entity. Flagged records are held for review.
Our team of CRE research analysts review and verify data daily. They cross-check each transaction against deal participants and property records. Data is never estimated or blindly added to the system.
Underwrite on verified comps, in the tools your team already uses.
Get StartedNet effective rent and concessions on the lease side; sale price, price per square foot, cap rate, NOI, and occupancy on the sale side; and loan amount, LTV, and maturity on the debt side.
Search the market directly rather than waiting on broker calls. Agent Connect returns statistics, trends, and the underlying records from a question asked in your own words.
Every comp is processed by a multi-step data verification system, including machine learning algorithms, statistical anomaly detection, and a team of CRE data analysts. Read more about our data standard.
Rent Predictor, which tests a rent assumption before it reaches the model, and Agent Connect, which runs underwriting questions inside Claude, ChatGPT, Gemini, or Copilot.
CompStak has collected 4M+ versions of lease and sale comps across 105 US markets, contributed by more than 40,000 verified CRE professionals and reviewed by CompStak’s analyst team.