Live workflow demo
AI Realtor Copilot - Live Demo
Search property listings, ask follow-up questions, and watch the structured panels update as the workflow progresses.
No listings yet.
Ask the assistant to search for homes, or start with one of the prompt examples below.
No listing selected yet.
After a search, ask about a specific result, such as Tell me more about the 2nd property.
Copy a scenario into chat, then watch the listing panels update.
Search
Do you have any places in Austin with at least two bedrooms and two bathrooms?
Do you have any 2 bed places in Houston for under $800K?
Refine
Are any close to a pool?
Can you narrow that down to homes with good commute options?
Inspect
Tell me more about the 2nd property in the list.
Tell me about property AUS-1-1.0-6.
Multi-step / agent routing
Do you have any 2 bed places in Houston for under $800K? Also, what is the typical home buying timeline? Also, what about mortgage pre-approval?
What information did you use to make that recommendation?
What should I ask next?
- Multi-step AI workflow orchestration.
- Conversational search with follow-up context.
- Structured listing results rendered outside the chat.
- Dynamic UI updates from agent state.
- A user-facing experience designed to feel transparent rather than opaque.
Built with LangGraph, CopilotKit, FastAPI, and Next.js.
Typical workflow
- 1Ask for homes that match a buyer's criteria.
- 2Refine the search with follow-up constraints.
- 3Inspect a specific listing.
- 4Use the structured UI to review results outside the chat.
Want something like this for your business?
This demo shows the kind of guided, production-aware AI workflow Ian builds for real teams.