OpenAPI Specs & llms.txt
Everything RealtyAPI exposes is machine-readable. Each provider publishes a public OpenAPI 3.1 spec, and a site-wide llms.txt indexes every API and endpoint for LLM agents. Both are public — no API key needed to read them.
OpenAPI 3.1 specs
Every provider serves its full spec — endpoints, parameters, and response schemas — from its own base URL:
https://{provider}.realtyapi.io/openapi.jsonFor example:
What to do with a spec
Generate typed clients
Feed the spec to openapi-generator, orval, or openapi-typescript for a fully typed SDK in your language.
Import into API tools
Postman, Insomnia, and Bruno import the spec directly — every endpoint pre-configured.
Power custom GPTs & actions
Paste the spec URL as an Action schema so ChatGPT (or any OpenAPI-aware agent) can call the provider.
Point coding agents at it
Agents read the spec and write the integration for you — parameters, paths, and response shapes included.
llms.txt
llms.txt is a convention for making a site legible to language models. Ours indexes all providers and 400+ endpoints — each with its method, path, and a link to full documentation:
https://www.realtyapi.io/llms.txt
Paste that URL into your agent's context (or just mention it in a prompt) and the agent can navigate our entire catalog on its own — pick the provider, find the endpoint, and read its docs page before writing code.
The one-paragraph integration prompt
No spec tooling at hand? This snippet is all an AI agent needs to wire up RealtyAPI:
You can call the RealtyAPI real estate data API.
Base URL: https://{provider}.realtyapi.io
(provider = redfin, airbnb, rightmove, zoopla, ...)
Auth: header x-realtyapi-key: YOUR_API_KEY
Format: every response is JSON
Index: https://www.realtyapi.io/llms.txt (all APIs & endpoints)
OpenAPI: https://{provider}.realtyapi.io/openapi.json (3.1)
Example:
GET https://redfin.realtyapi.io/properties/search?location=Seattle,WAWant live tool-calling instead of code generation? Use the MCP server. Working in Claude Code? The Agent Skill bundles all of this knowledge into one install.