Skip to content
Start free trial
Pricing Start free trial

Trust · Fair Housing

Fair Housing compliance, built in.

Generic AI does not know Fair Housing and will write copy that gets an agent fined. RealtrAI puts a screen between the model and publish, on every client-facing output, governed by Trunnion AI.

Reviewed by The RealtrAI Fair Housing desk · Last reviewed

Start free trial Security & governance
RealtrAI Fair Housing screen Live Live
Flagged

"Perfect for a young Christian family near the church."

Steering · references familial status, religion

Passed

"Perfectly sized for entertaining, with a welcoming layout and a vibrant, walkable neighborhood."

Describes the property, not the buyer

  • 01 Pre-generation filterProhibited language patterns blocked at input
  • 02 HUD guideline reviewDraft checked before it reaches your editor
  • 03 Audit logEvery flag, override, and justification recorded
Screened against
RaceColorNational originReligionSexFamilial statusDisability + state & local

Our commitment

Non-compliant content should never auto-publish.

The Fair Housing Act prohibits discrimination in housing, and that includes the words in a listing. AI makes it faster than ever to write those words, which means it makes it faster than ever to write the wrong ones. RealtrAI is built so the speed of AI never becomes a compliance liability: every client-facing draft is screened, logged, and reviewed before it goes out.

Protected classes

The seven federal classes, plus the local ones.

RealtrAI screens against the seven protected classes under the federal Fair Housing Act, and accounts for the additional classes many states and cities add.

RaceColorNational originReligionSex (including sexual orientation and gender identity)Familial statusDisability

Common state & local additions

Source of incomeAgeMarital statusCitizenship / immigration statusVeteran or military statusGenetic informationDomestic violence survivor status

Platform safeguards

A three-tier screen, plus a human.

01

Pre-generation filter

Inputs that contain prohibited language patterns are flagged before any output is generated, so the model never builds on a non-compliant prompt.

02

HUD-guideline output review

Every client-facing draft is reviewed against HUD advertising guidelines before it reaches your editor, with the specific risk called out.

03

Audit log

Every flag, override, and override justification is captured in an audit log, so a brokerage can document compliance and review patterns over time.

04

Human-in-the-loop

Nothing auto-publishes. A person reviews and approves the output, keeping the agent responsible and in control.

Steering & discriminatory language

Describe the property, not the buyer.

The most common Fair Housing slip in marketing copy is language that describes who should live somewhere instead of what the home is. RealtrAI flags steering and protected-class references, then rewrites to describe the property and the place, so the listing stays compliant without losing its voice.

See it in Listing Writer
RealtrAI Fair Housing screen Live Live
Flagged

"Perfect for a young Christian family near the church."

Steering · references familial status, religion

Passed

"Perfectly sized for entertaining, with a welcoming layout and a vibrant, walkable neighborhood."

Describes the property, not the buyer

  • 01 Pre-generation filterProhibited language patterns blocked at input
  • 02 HUD guideline reviewDraft checked before it reaches your editor
  • 03 Audit logEvery flag, override, and justification recorded
Screened against
RaceColorNational originReligionSexFamilial statusDisability + state & local

What the tool does not replace

You stay responsible. We make it easy to get it right.

  • i Your professional judgment and final review of every word that goes out under your name.
  • i Your brokerage compliance policy, your MLS rules, and your state and local requirements.
  • i Legal advice. RealtrAI helps you avoid common Fair Housing pitfalls; it is not a law firm.

Reporting & resources

Go to the source.

RealtrAI's screening follows federal Fair Housing law and HUD advertising guidance. The primary sources, so you can verify the rules yourself:

To report a suspected Fair Housing violation, contact HUD or your state or local fair-housing agency. RealtrAI's screen is a safeguard built into the product, not legal advice.

FAQ

Fair Housing, answered.

What is Fair-Housing-safe AI?

Fair-Housing-safe AI is generative AI with a compliance screen between the model and publish. RealtrAI screens every client-facing output against the seven federal protected classes and HUD advertising guidelines, keeps an audit log, and requires human review before anything is published, so AI speed does not become Fair Housing risk.

Which protected classes does RealtrAI screen for?

The seven federal protected classes under the Fair Housing Act: race, color, national origin, religion, sex including sexual orientation and gender identity, familial status, and disability. Many state and local laws add more, such as source of income, age, and marital status, and RealtrAI accounts for common additions.

Does RealtrAI guarantee compliance?

No tool can guarantee compliance, because the agent is responsible for what they publish. RealtrAI dramatically lowers risk by filtering prohibited language, reviewing against HUD guidelines, logging every decision, and keeping a human in the loop. You stay in control and you keep the audit trail.

How does this help a brokerage?

A brokerage gets consistent Fair-Housing-screened output across every agent, plus an audit log of every flag and override for compliance documentation. On the Brokerage plan, that pairs with SSO, per-tenant isolation, and SOC 2 documentation. See /security.

Get started

AI you can publish with confidence.

Try the Fair Housing screen on your own listings, free for seven days.

Start free trial Explore the tools