Free Resource

The FlexMLS + ChatGPT Prompt Pack

You connected FlexMLS to ChatGPT. Now what do you actually type? Here are 22 prompts that work — market snapshots, listing searches, CMA prep, and content you can publish. Copy one, swap the highlighted parts for your own ZIP codes and price points, and go.

Tennessee REALTORS® · Updated August 2026 · TREC-approved CE at Home Star Courses

Jump to the prompts Watch the video

Watch first

See the FlexMLS connector in action

This walkthrough shows exactly how to connect FlexMLS to ChatGPT and what happens when you run these prompts against live MLS data.

Haven't connected it yet?

Three steps, about five minutes

The full walkthrough is in the video. Here's the short version so this page stands on its own.

1

Turn it on in FlexMLS

Menu → scroll to AI Settings → switch the MCP Server from disabled to enabled.

2

Copy the server URL

Click MCP Server Setup and copy the address: https://mcp.flexmls.com/mcp

3

Add it in ChatGPT

Plugins → + → name it FlexMLS, paste the URL, choose OAuth, then sign in with your FlexMLS credentials.

Read this first

Three things to know before you start

You need a paid ChatGPT plan

Custom connectors require a Pro or Plus account — roughly $20 a month. If you're not already using AI daily, that may not be worth it yet.

It reads. It doesn't write.

The connector can search and analyze your MLS data. It cannot change a listing, edit a record, or push anything back into FlexMLS. That's why it's safe to explore.

Check the numbers before you publish

Treat every figure as a draft. Open FlexMLS, confirm the numbers make sense, and only then put it in front of a client. Build the trust first.

The prompts

Copy, swap the yellow bits, send

Anything highlighted is a placeholder — replace it with your own ZIP code, price range, or address. The more specific you are, the better the answer. Garbage in, garbage out.

★

Market snapshots

5 prompts

Start here. These are the fastest way to see whether the connection is pulling real data.

The one from the video

Give me a 1,000-word blog post on what has been happening in the market in the 37934 ZIP code over the last year.

Fastest sanity check

Summarize the market statistics for my MLS over the last month.

Head-to-head comparison

Compare median sale price, average days on market, and months of inventory for ZIP A versus ZIP B over the last 12 months. Put it in a table and tell me in one paragraph which area is moving faster and why.

Trend over time

For ZIP code, show me how average days on market has changed month over month for the last 12 months. Then tell me what that trend means for a seller who is pricing a home this week.

Sale-to-list ratio

What percentage of original list price did homes in ZIP code sell for in each of the last 12 months? Flag any month that was unusually high or low and explain what was happening.
★

Finding and filtering listings

4 prompts

Plain-English searches instead of checkboxes. Be specific about property type, price band, and timeframe.

Your own inventory

Show me my active listings, sorted by days on market, with the current price and the date of the last price change.

Stale inventory hunt

Find single-family homes in ZIP code between $300,000 and $500,000 with at least 3 bedrooms that have been on the market more than 60 days. Show address, list price, days on market, and any price reductions.

Price-drop watch

List every residential property in area or ZIP that had a price reduction in the last 30 days. Show original list price, current price, dollar change, and percent change, sorted by largest reduction.

Where the competition isn't

In ZIP code, break the market into $50,000 price brackets. For each bracket, show me the number of active listings versus the number of closed sales in the last 90 days, and tell me which bracket has the least competition for a seller right now.
★

Pricing and CMA prep

4 prompts

These save the most time. Always open FlexMLS and verify the comps before they reach a seller.

Comp pull

Pull the last 6 closed sales in subdivision or ZIP that are within 15% of 2,200 square feet and built within 10 years of 1998. Show address, close price, price per square foot, days on market, and close date.

Defensible price range

For property address, find comparable closed sales from the last 6 months and give me a defensible list price range. Show your reasoning and name the specific comps you leaned on most.

The cost of overpricing

In ZIP code, compare homes that sold in under 30 days against homes that took more than 90 days. For each group show the average sale-to-original-list-price ratio and the average number of price reductions. Write me two sentences I can say to a seller who wants to list high.

Seasonality

Over the last three years in ZIP code, which months had the highest number of closings and the strongest sale-to-list ratios? Tell me the best month to launch a listing and back it with the numbers.
★

Turning data into content

5 prompts

This is where the time savings compound. One search becomes a blog post, a reel, a caption, and an email.

SEO blog post

Using current MLS data, write an 800-word blog post about the city or ZIP market for buyers relocating to the area. Include one H1, three H2 sections, and a meta description under 160 characters. Write for a homeowner, not a statistician.

Instagram carousel

Turn the last quarter of market data for ZIP code into an Instagram carousel: a 4-slide outline with one statistic per slide, a caption under 150 words, and 8 relevant hashtags. Keep it conversational.

60-second video script

Write a 60-second video script I can read straight to camera about what is happening in the ZIP code market right now. Conversational, no jargon, one clear takeaway, and a natural call to action at the end.

Email to your sphere

Write a short email to my past clients about the ZIP code market this quarter. Under 200 words, friendly, one interesting number, and end with an invitation to reply if they're curious what their home is worth.

Listing presentation sheet

Using current data for ZIP code, write me a one-page listing presentation talking-point sheet: five bullet points a seller needs to hear, each backed by a specific number, plus one sentence explaining what that number means for them.
★

Buyer and seller advisory

3 prompts

Answering the questions clients actually ask, in language they actually understand.

Whose market is it?

Based on current inventory and absorption rate in ZIP code, is this a buyer's market or a seller's market? Give me the months-of-inventory figure, then explain what it means in plain English to someone who has never bought a house.

What the budget actually buys

For a buyer with a budget of $425,000 in area, what does that realistically get them right now? Summarize typical square footage, bedroom count, year built, and lot size, and name two neighborhoods worth looking at.

Negotiating room

In ZIP code over the last 90 days, how far below list price are buyers actually getting homes? Break it out by price band and tell me where a buyer has the most leverage.
★

Checking its work

1 prompt

Use this one constantly, especially in your first few weeks. It is the difference between a tool you trust and a tool that embarrasses you.

Make it show its sources

Show me exactly which MLS fields, property types, and date range you used for that answer, and list the specific listings it was based on, so I can verify it in FlexMLS myself.

Tips and gotchas

How to get better answers

Treat it like a new assistant

Give it a broad instruction and you'll get a broad answer. Name the ZIP code, the property type, the price band, and the timeframe every single time.

Garbage in, garbage out.

Watch the Gmail crossover

If you have Gmail connected to ChatGPT and you ask for a "saved search," it will sometimes go looking in your inbox instead of your MLS — because that's where your listing notifications live.

Say "using the FlexMLS connector" in the prompt to keep it honest.

Ask follow-up questions

The first answer is a starting point, not the deliverable. Push back. Ask it to narrow the date range, exclude new construction, or redo the analysis by price band.

Give it a job, not just a question

"What's the market doing in 37934?" gets you a summary. "Write five listing-presentation bullets for a seller in 37934, each backed by a number" gets you something you can use this afternoon.

One search, many pieces of content

Once it has pulled the data, don't start over. Say "now turn that into an Instagram caption," then "now a video script," then "now an email." It keeps the same numbers throughout.

Verify before anything goes public

Open FlexMLS and confirm the figures before a blog post, a listing presentation, or a client email goes out. The tool is fast. Your reputation is slow to rebuild.

Use it like a professional

The rules that keep you compliant

  • You verify, always. AI summarizes MLS data — it doesn't take responsibility for it. Before a number reaches a client or an ad, check it against the source record. Accuracy is your license-law duty.
  • Fair housing applies to prompts too. Search and describe properties and amenities — never people or who "belongs" in a neighborhood. If a prompt asks about demographics, rewrite the prompt.
  • Cite your source and date. "Per our MLS, this month's data" — a confident wrong number in a market update lives forever.
  • Follow your MLS's data rules. How you display, share, and advertise MLS data is governed by your MLS agreement. The connector doesn't change that.

Want to go deeper?

Home Star Courses teaches TREC-approved continuing education for Tennessee REALTORS® — including live classes in AI, technology, and video production. Every hour counts toward your 10 elective hours.

See upcoming classes