Review · Updated October 2026
HomeSage AI Review 2026: Pricing, Credits and Whether It’s Worth It
A plain-English look at the AI real estate data platform for investors, agents, lenders and developers: what each credit buys and where the costs hide.
By Oyekale Olawale · Last verified October 26, 2026
Quick Answer
HomeSage AI (Homesage.ai) is an AI real estate data platform. It scores investment deals, builds Full Property Reports and serves data on 155M+ US properties through a web app, 31 REST endpoints and an MCP server.
Plans start at $100/month (Starter, 750 credits). The BETA25 coupon drops that to $75 until October 31, 2026. A Full Property Report costs 15 credits, so Starter covers about 50 of them.
My verdict: worth it if you screen deals every week or build products on US property data. Skip it if you analyze a handful of properties a year.
Test HomeSage AI with 300 free credits
Every self-serve plan starts with a 30-day trial: 300 credits, $0 today. That’s enough for 20 Full Property Reports.
Start the free trialWhat HomeSage AI Actually Does
HomeSage AI is built for people who make money on property decisions. You give it a US address, or a set of filters, and it returns value, comps, renovation cost, rental income and investment returns.
The same data comes in four forms: a web app with Investment Property Search and Full Property Reports, an AI agent called Sage, a REST API, and an MCP server for assistants like Claude and ChatGPT. One credit balance pays for all of them.
| Product | What it does | Best for |
|---|---|---|
| Sage | Chat-style agent that returns a verdict, comps, rehab cost, rent and owner data | Quick underwriting |
| Investment Property Search | Ranks on-market and off-market deals against your filters | Finding deals |
| Full Property Report | 292 data fields from 11 sources, with what-if scenarios | Due diligence |
| Real Estate APIs | 31 REST endpoints across five categories | Developers, proptech |
| MCP server | 33 tools inside Claude, ChatGPT, Gemini, Perplexity and Cursor | AI-assistant workflows |
| DealFinder | Chrome extension plus iOS and Android app for checking listings as you browse | Field and browsing use |
There’s also a Custom AI Solutions arm that sells an AI strategy assessment, from $350 for companies with 1–10 employees to $750 for 100+.
How I Evaluated HomeSage AI
Here’s the honest scope. I went through the live pricing page, the API library, the MCP documentation, the Sage demo and the company’s own announcements. Then I ran the credit math myself and checked plan prices against G2 and Capterra listings.
What I did not do is audit its valuations against closed sales. No pricing-page review can. Treat every value and ARV figure as an estimate. Where HomeSage’s own pages disagree with each other, I point it out below.
HomeSage AI Pricing 2026: Every Plan and the Real Cost per Report
HomeSage AI is credit-based. Each plan gives you a monthly credit balance, and every report or API call spends some of it. Yearly billing takes 20% off the monthly price.
| Plan | List → BETA25 | Credits | Per credit | Full reports | Per report |
|---|---|---|---|---|---|
| Starter | $100 → $75 | 750 | $0.13 | ~50 | $2.00 |
| Small | $200 → $150 | 1,250 | $0.16 | ~83 | $2.40 |
| Medium | $500 → $375 | 5,000 | $0.10 | ~333 | $1.50 |
| Large | $1,000 → $750 | 20,000 | $0.05 | ~1,333 | $0.75 |
| Enterprise | Custom | Custom | Negotiated | Custom | Negotiated |
“Per report” uses list prices and the 15-credit Full Property Report. With BETA25 the cost per report falls to about $1.50 (Starter), $1.80 (Small), $1.13 (Medium) and $0.56 (Large).
The pricing page marks the 25% coupon as “locked-in pricing” and says the offer runs until October 31, 2026, with “other restrictions may apply.” Screenshot your checkout so you have the rate on record. Billed yearly at list price, the four plans work out to $80, $160, $400 and $800 a month. The MCP page shows Starter at $60 a month with yearly billing plus BETA25.
The Small plan is the pricing quirk
Starter gives you 750 credits for $100. Small gives you 1,250 for $200. That extra 500 credits costs $100, or $0.20 each, which is worse than Starter’s $0.13 and double Medium’s $0.10.
My read: start on Starter, and jump to Medium when you outgrow it. I couldn’t find a top-up price on the pricing page, so ask support if you need something between the two.
Run your own numbers before you pay
Use the 300 trial credits on the three or four properties you’re already evaluating.
Claim 300 trial creditsWhat Each Credit Buys
Every endpoint has a fixed credit price. These are the costs listed on the pricing page today:
| Endpoint | Credits | Returns |
|---|---|---|
| Full Property Report | 15 | Full report with proprietary investment indicators |
| Long-Term Rental Estimates | 5 | Gross and net income, IRR, cap rate, NOI, cash-on-cash |
| Short-Term Rental Estimates | 5 | Same metrics for short-term rental use |
| Renovation Cost | 5 | Rehab cost, full potential, spread, value gap |
| Comps | 5 | Comparable sales with adjusted prices |
| Property Condition | 3 | Excellent, Good, Outdated, Poor or Very Poor |
| Investment Potential | 3 | Good, High or Excellent equity potential |
| Renovation Return | 3 | Project cost, profit, ROI, max recommended budget |
| Flip Return | 3 | Resale ROI, profit, project cost |
| Seller Motivation Score | 2 | Negotiation signal (the PFS value in the response) |
| Property Info | 2 | Listing details, days on market, history, school ratings |
| TLC Indicator | 1 | Yes/No flag for renovation potential |
| Home Value Graph | 1 | Home value history |
| Bulk Property Info | 1 per property | Condition, ROI, profit and costs for many properties at once |
| Auto Complete | 1 (see note) | Address suggestions and property IDs |
The Auto Complete row is one of the inconsistencies I found. The pricing table lists 1 credit, but the API library and the MCP page both call address autocomplete free. Check your dashboard before you build a feature on it.
Screen first, then spend 15 credits
Buying the pieces separately is the expensive route. Property Info (2) + Comps (5) + Renovation Cost (5) + Flip Return (3) + Long-Term Rental (5) adds up to 20 credits. The Full Property Report covers that same ground for 15, based on what the report shows: ARV, comps, rehab cost, and flip and rental returns.
The smarter move is a cheap screen. Investment Potential (3) + TLC (1) + Seller Motivation Score (2) costs 6 credits and tells you whether a listing deserves the full report.
The Features That Matter
Investment Property Search
You filter by Investment Potential, Seller Motivation Score, Property Condition, Value Gap, location, listing price and days on market. Results are ranked by potential, not by listing date, and cover both MLS and off-market properties.
Saved searches and daily updates are included. One cost warning: database search runs 1 credit per matching property, so a loose filter can burn credits fast.
Full Property Report
One report pulls 292 data fields from 11 sources: value, condition, comps, returns, schools and neighborhood risk. The reports are interactive, so you can change assumptions and watch profit and ROI move.
Sage, the AI agent
Sage answers plain-English questions about an address with a verdict (PASS, WATCH, BUY or STRONG BUY), ARV, comps, rehab cost, rent and owner details. It shows its steps, and it states a range and a confidence label instead of one confident number.
Three extras stood out. You can attach walkthrough photos for a room-by-room rehab budget. You can upload a CSV or Excel list and have it add value, rent and owner data. And a deep-research mode writes a ranked report with a spending cap, so it can’t run away with your credits.
Condition scoring from listing photos
Computer vision grades condition on a five-step scale, from Excellent to Very Poor, for 3 credits. HomeSage says it was first to run computer vision on all new listings. That’s a vendor claim, but the output is useful: condition drives the rehab estimate, which drives the flip math.
Skip tracing, liens and owner data
Five skip-tracing endpoints return owner names, mailing addresses, phone numbers and emails, plus associated people. Bulk versions handle up to 100 properties per call. Mortgage and lien lookups are cached free for seven days.
It’s a different animal from the consumer people-search sites you may have used, because the data is tied to a property record. Contact data is only useful if you reach people lawfully, so check TCPA and state rules before you dial or text.
Using HomeSage AI Inside Claude or ChatGPT (MCP)
The MCP server is the newest piece, and the most interesting. It lets an assistant call HomeSage data directly, so you can type “Analyze [address]” and get a decision-ready answer without leaving the chat.
- Copy your personal connection link from the developer dashboard.
- Add it as a custom connector in your assistant’s settings. In Claude on web or desktop, that’s Customize → Connectors → the “+” → Add custom connector.
- Approve the HomeSage sign-in, then ask about any US address. On Claude Team or Enterprise, an admin adds it in organization settings first.
Paid tools spend credits, and your assistant confirms the cost before charging. Address and area autocomplete, search previews and your credit balance are free.
One rough edge: the MCP page says “no code, no keys” in the setup steps, but its FAQ says you connect “with your API key.” Expect the sign-in flow, and check what your dashboard shows.
Connecting any outside data source to an assistant is also a trust decision. I covered how far to trust Claude’s answers in my piece on whether Claude is safe and trustworthy.
For Developers: API Details, Limits and Costs
The API is a Bearer-token REST API with a Developer Playground and copy-paste cURL, JavaScript and Python examples. A first call is one request. For example, the current-estimate endpoint on developers.homesage.ai takes a property_address parameter and costs 2 credits.
The cost mechanics worth knowing before you build:
- Failed API calls don’t cost credits, and a few endpoints are free, according to the company’s docs.
- Bulk Property Info charges 1 credit for each property returned, not per request.
- Mortgage and lien lookups are cached free for seven days.
- Data refreshes daily, and the vendor claims 99.9% uptime and sub-100ms responses. I couldn’t verify either independently.
Proptech startups can also use a free sandbox: sign up, claim 300 credits and start calling the API. When the credits run out, the next step is Starter at $100 a month.
If you’re prototyping a front end, an AI app builder like Famous.ai or Macaly gets you a screen fast. For a no-code build, read how Bubble’s workload units are priced first, because API-heavy apps can surprise you there too.
HomeSage AI vs PropStream vs RentCast
These three overlap less than they look. PropStream is built for lead lists and outreach. RentCast is a cheap rent and value data API. HomeSage sits in the middle, with deal underwriting plus an API and an MCP server.
| HomeSage AI | PropStream | RentCast API | |
|---|---|---|---|
| Best for | Deal analysis, API and AI-assistant access | Lead lists, skip tracing, outreach | Rent and value data in your own app |
| Entry price | $100/mo (Starter) | $99/mo (Essentials) | $74/mo (Foundation) |
| Free option | 30-day trial, 300 credits | 7-day trial, 50 free leads | Free plan, 50 requests/mo |
| Pricing model | Credits per report or call | Tiers by monthly saves and exports | Request quota plus overage fee |
| Coverage | 155M+ US properties | 160M+ US properties | Nationwide US datasets |
| Standout | Seller Motivation Score, official MCP, Sage | Skip tracing free on Pro and Elite | Low-cost entry for developers |
Prices are the entry plans as of October 2026. Units differ, so don’t compare them line for line: one RentCast request is not one HomeSage credit. If you want to see how PropStream behaves in a title-and-lien workflow, I tested it in my AI title search roundup.
Pros, Cons and Rough Edges
- ✓Full report (15 credits) is cheaper than buying the pieces (20)
- ✓One credit balance across app, API, MCP and Sage
- ✓30-day trial with 300 credits
- ✓Per-endpoint credit prices are published
- ✓Official MCP server with 33 tools
- ✓Failed API calls cost nothing
- ✗Small costs more per credit than Starter
- ✗Database search burns 1 credit per matching property
- ✗US properties only
- ✗Values and ARV are estimates, not appraisals
- ✗Auto Complete is 1 credit in one place and free in two others
- ✗MCP setup says “no keys” and “API key” on the same page
HomeSage is upfront about the estimate problem. Sage’s own FAQ says a verdict is a starting point: check the comps, walk the property and order title before you buy or lend. Take that seriously.
Who Should Use HomeSage AI
Agents and brokers can use it for pricing a listing, prepping investor clients and sizing up a fixer-upper. If you’re still choosing your stack, my roundup of AI tools for real estate agents covers the wider field.
Skip it if you’re a first-time buyer, you analyze fewer than ten properties a year, or you work outside the US. If you’re new to investing and not buying rentals or flips yet, beginner investing apps are a better first step.
Check the BETA25 price before it ends
Starter is $75 a month with the coupon, per the pricing page, through October 31, 2026.
See HomeSage AI plansFAQ
What does HomeSage AI do?
HomeSage AI analyzes US residential properties and returns values, comps, renovation costs, rental income and investment returns. You can use it through a web app, an AI agent called Sage, a REST API or an MCP server.
How much does HomeSage AI cost?
Starter is $100 a month for 750 credits, Small is $200 for 1,250, Medium is $500 for 5,000 and Large is $1,000 for 20,000. Enterprise is custom. Yearly billing takes 20% off, and the BETA25 coupon takes 25% off through October 31, 2026.
Is there a free trial?
Yes. Every self-serve plan starts with a 30-day trial that includes 300 credits and costs $0 today. Some third-party listings still quote older trial details, so trust the pricing page.
How many credits is a Full Property Report?
15 credits. That’s about 50 reports on Starter, 83 on Small, 333 on Medium and 1,333 on Large.
Does HomeSage AI work with Claude and ChatGPT?
Yes, through its MCP server. It works in Claude, ChatGPT, Gemini, Perplexity and Cursor. You add your connection link as a custom connector, sign in once, then ask about any US address.
Which areas does it cover?
HomeSage AI covers 155M+ US properties across all 50 states. It doesn’t cover properties outside the US, and coverage varies by area.
How accurate are its valuations?
They’re estimates. Sage shows a range and a confidence label, and the company tells users to check comps, inspect the property and order title before buying or lending.
Is HomeSage AI better than PropStream?
They solve different problems. PropStream is stronger for lead lists and outreach. HomeSage is stronger for deal underwriting and for developers who want an API or MCP connection.
Conclusion
HomeSage AI earns its place if you analyze properties often. The credit model is easy to reason about, the 15-credit report beats buying the pieces separately, and one balance covering the app, API and MCP is a real convenience.
The weak spots are small but real: a Small plan that’s worse value than Starter, a few inconsistencies between its own pages, and estimates you still have to verify. None of them is a dealbreaker.
My advice is simple. Use the 300 trial credits on properties you’re already weighing, compare the numbers against your own comps, and only then pick a plan. Start on Starter, and move to Medium when you outgrow it.
Disclosure: this article contains affiliate links. If you sign up through them, I may earn a commission at no extra cost to you. Prices and offers were checked on October 26, 2026 and can change, so confirm them on the pricing page.
