Casafari MCP: European property data for AI agents
Casafari MCP gives AI assistants and coding tools direct access to CASAFARI's deduplicated European property graph across 16 countries (US: entering now, selected states, commercial real estate only, via Claro AI). Available on platform.casafari.com through five specialized Model Context Protocol servers—Location Lookup, Area Insights, Comparables & Valuation, AgentGraph, and Claro—the service authorizes seamlessly via OAuth in your browser without copying API keys.
Casafari MCP gives AI assistants and coding tools direct access to CASAFARI's deduplicated European property graph across 16 countries. Available on platform.casafari.com through five specialized Model Context Protocol servers—Location Lookup, Area Insights, Comparables & Valuation, AgentGraph, and Claro—the service authorizes seamlessly via OAuth in your browser without copying API keys. A Free plan offers 1,000 usage tokens for 30 days with no credit card required.
What is Casafari MCP and who is it for?
Imagine an acquisitions manager in London sitting down on a Friday afternoon to evaluate expansion options for Spain. Instead of manually cross-referencing multiple portals, she opens Claude and asks: "How many empty restaurant premises between 120 and 200 m² are currently available to let in Madrid's Salamanca district?" Rather than offering a generalized guess, the AI assistant calls a dedicated server and returns an exact count backed by verified market listings, extracted criteria, and local footfall data.
This capability is built on the Model Context Protocol, an open standard that standardises how large language models connect to live data sources. Behind Casafari MCP lies CASAFARI's clean property graph covering residential and commercial properties, for sale and for rent, across 16 countries (CASAFARI developer docs, how the property graph is built, October 2026). Every property is consolidated into a single master record featuring full price history and status logs.
Casafari MCP is designed for three distinct workflows:
- Software developers who need to integrate real estate intelligence into custom workflows or coding assistants within five minutes.
- Data and research teams using tools like Cursor or Claude Code to run automated market analyses and property trends inside notebooks.
- Real estate professionals, investors, and expansion managers who prefer asking natural-language questions in Claude or ChatGPT and receiving structured answers grounded in verified market facts.
Which MCP servers can I connect?
CASAFARI provides five distinct MCP servers on platform.casafari.com. You can connect only the servers required for your specific workflow. Location Lookup acts as the foundational zero-token resolver that automatically translates place names into system identifiers before other tools execute.
Location Lookup
Location Lookup resolves place names, districts, and municipalities into standard `location_id` parameters used across all other CASAFARI servers. Calls to Location Lookup cost 0 usage tokens. Your AI assistant executes this lookup transparently whenever you mention a geographical area. `https://platform.casafari.com/mcp/locations`
- "Which location id covers Chamberí in Madrid?"
- "Find the parish of Estrela in Lisbon."
- "Is there a district called Prenzlauer Berg in Berlin?"
Area Insights
Equipped with 7 specialized tools, the Area Insights server delivers aggregate market statistics, temporal price trends, rental price distributions, and geographic heatmaps. To learn more about standard analytical capabilities, visit Area Insights, CASAFARI's market analytics product or read the Area Insights developer docs. `https://platform.casafari.com/mcp/area-insights`
- "How have asking prices per m² for flats moved in Porto over the last three years?"
- "Show me the rent distribution for two-bedroom flats in Valencia."
- "Which parts of Milan have the highest asking prices per m² right now?"
Comparables & Valuation
This server evaluates comparable active and historical listings around specific coordinates, physical street addresses, or Spanish cadastral references (`referencia catastral`). If you work with Spanish property identifiers, read our guide on how a Spanish cadastral reference helps a valuation. For further integration details, check the Comparables and valuation developer docs and explore CASAFARI's valuation and market study tools. Note that estimated price outputs are data-driven statistical benchmarks rather than formal survey valuations. `https://platform.casafari.com/mcp/comparables`
- "What are the closest comparables for a 95 m² flat at this address in Barcelona?"
- "Estimate a price range for cadastral reference 1234567VK4713S0001AB."
- "What sold or was withdrawn within 500 m of this point in the last 12 months?"
AgentGraph
AgentGraph provides access to 13 specialized tools mapping real estate agencies and individual brokers across Germany. It provides verified business data to help firms recruit top talent or establish local co-brokerage partnerships. Technical specifications can be reviewed in the AgentGraph developer docs. `https://platform.casafari.com/mcp/agentgraph`
- "Which agencies in Munich list the most family houses above EUR 1.5 million?"
- "Who are the most active agents for rentals in Hamburg-Eimsbüttel?"
- "Which agencies in Frankfurt have grown their listing stock fastest this year?"
Claro
Claro is CASAFARI's dedicated business expansion search engine, featuring 9 MCP tools for commercial site selection. It screens shops, restaurant spaces, offices, and warehouses across Spain and Germany, as well as commercial plots in Germany. The same engine powers the Claro app at expansion.casafari.com and CASAFARI for Business Expansion. For API reference details, see the Claro documentation. Every response presents exact match counts where criteria are explicitly categorized as met, not met, or not stated in the original listing, complete with source quotes. Additionally, Location Intelligence features provide contextual data including a footfall index within a 5-minute walk, local resident counts, and purchasing power figures. `https://platform.casafari.com/mcp/claro`
- "Find ground-floor premises of 80-150 m² with extraction for a coffee bar in Valencia."
- "Which of these Barcelona premises ask for key money?"
- "Find commercial plots of 1,500-6,000 m² near motorway junctions in North Rhine-Westphalia for a vehicle inspection centre."
How do I connect Casafari MCP to Claude, ChatGPT, Gemini CLI, Claude Code or Cursor?
Connecting an AI assistant takes under two minutes through standard Streamable HTTP transports:
Copy the URL of the server …
Copy the URL of the server you wish to connect from platform.casafari.com.
Open your client's settings
Open your client's settings, add a new MCP server or custom connector, and paste the endpoint URL.
Complete the OAuth 2.1 sign-in in your browser window
Complete the OAuth 2.1 sign-in in your browser window. No API keys are required.
Configuration locations vary slightly by client:
- Claude (claude.ai, Desktop, mobile): Navigate to Settings > Connectors and add a custom connector.
- ChatGPT: Enable Developer Mode and configure a custom connector.
- Claude Code & Gemini CLI: Add the server URL directly via the command line interface.
- Cursor: Add the endpoint under Cursor Settings > MCP.
Casafari MCP supports any client implementing the Model Context Protocol, including VS Code, Codex CLI, Devin Desktop (formerly Windsurf), Visual Studio, Zed, Cline, JetBrains Junie, Kiro, Warp, and goose (CASAFARI developer docs, connection guide with snippets for every client on developers.casafari.com, October 2026).
What does an AI assistant actually do with it?
By combining reasoning capabilities with live data access, AI agents can screen thousands of market listings down to actionable shortlists in seconds.
In October 2026, a test expansion report was executed using Claude paired with the Claro connector. Claude screened 2,296 commercial premises across Madrid and Barcelona for a smash-burger brand, shortlisting six sites that fully satisfied technical extraction and size requirements while detailing key-money obligations (CASAFARI developer docs, example reports written by Claude with the Claro connector, October 2026; view the open files on the public casafari-mcp repository with screenshots and example reports).
Multiple servers can also be queried in a single conversation. An agent can use Claro to identify commercial spaces in Catalonia and immediately call Area Insights to evaluate neighborhood rental yields.
If you regularly use generative tools for property research, see our companion guide on ChatGPT prompts for estate agents.
What do the plans cost, and what is a token?
Casafari MCP measures tool consumption in usage tokens. These usage tokens are separate from the OAuth access tokens stored by your client.
- Free Plan: Includes 1,000 usage tokens valid for 30 days upon initial sign-in. No credit card is required.
- Agent Plan: EUR 49 per month billed annually (or EUR 70 month-to-month), excluding VAT.
- Custom Plan: Customized enterprise volumes and limits available by agreement.
Location Lookup queries consume 0 tokens. For what each tool costs, consult the CASAFARI developer docs on plans, tokens and what each tool costs. For enterprise agreements, see how the Custom plan works and request it from the platform.
How secure is it?
Casafari MCP complies with the Model Context Protocol specification for HTTP transports using OAuth 2.1 authentication. You sign in directly through your browser, eliminating the need to save static, long-lived API keys in local configuration files. You can review or revoke client permissions at any time inside your account profile.
The data tools across the five MCP servers operate in read-only mode; only AgentGraph's watchlist tools (save_to_watchlist and remove_from_watchlist) save or remove entries in your organisation's shared watchlist. AI assistants can retrieve market trends, listing information, and agency records, but cannot otherwise write, modify, or delete data within your account. Management and authentication settings are hosted securely on platform.casafari.com, where you can sign in to platform.casafari.com.
Already using the CASAFARI REST API?
If you currently consume property intelligence through the CASAFARI Property Data API at api.casafari.com, your setup remains completely unchanged. The REST API continues to serve programmatic software applications, while Casafari MCP provides a parallel interface optimized for AI tools. To see how MCP tools correspond to REST endpoints, consult the full tool catalogue for all five MCP servers.
To manage your active endpoints or start testing with a Free account, visit your account and servers on platform.casafari.com.
Testimonials
The APIs allow us to obtain information in real time, being able to quickly identify changes in the stock for sale. All this information is crucial in the process of monitoring the market, whether for intelligence purposes or for property acquisition transactions.
By having access to a database of useful and structured data for our AI engine, we were able to scale our real estate portal faster. The main advantages are the speed and updating of the data used, which allows us to keep real estate listings always synchronised with the market and exhibit our customers new listings on the platform every day.
We are at the threshold of 20,000 registered users in less than six months of operations, and certainly being able to have CASAFARI’s API to add listings to our portfolio was one of the reasons for this achievement.
Registered transaction prices from official sources
CASAFARI holds registered transaction prices: what homes actually sold for, taken from official sources in Spain, Portugal, France, Italy, Poland and Austria, and not only what they were advertised at. Your AI agent works from the same market data that powers CASAFARI's own tools.
- Comparables: sold homes with a real closing price carry a purple “closing price” tag and are added automatically when the valuation finds them.
- CMA: the estimate multiplies the average price per m² of your selected comparables by the property's area, so comparables with registered prices count towards the value.
- Market Study: see the average closing price for the kind of property you are valuing in a chosen area, next to current asking prices.
- Property History: every registered sale around the property in Spain, France, Italy and Poland, and how the average closing price in the area has developed. In Portugal and Austria it shows the development of aggregated official statistics for the area.
Official sources: Notariado in Spain, INE in Portugal, DVF in France, Agenzia delle Entrate in Italy, RCN in Poland and Statistik Austria in Austria.
Frequently asked questions
Is Casafari MCP free to try?
Yes. A Free plan is automatically created upon your first OAuth sign-in and provides 1,000 tokens valid for 30 days, with no credit card required (CASAFARI developer docs, October 2026). Location Lookup requests cost 0 tokens, allowing you to resolve place names without spending your token balance. If you need additional volume, the Agent plan is available for EUR 49 a month billed annually or EUR 70 month-to-month, ex VAT.
Which AI assistants and tools work with Casafari MCP?
Casafari MCP supports any client that implements the Model Context Protocol over Streamable HTTP. Confirmed clients include Claude (claude.ai, Desktop, mobile), ChatGPT in Developer mode, Claude Code, Cursor, VS Code, Codex CLI, Gemini CLI, Devin Desktop (formerly Windsurf), Visual Studio, Zed, Cline, JetBrains Junie, Kiro, Warp, and goose (CASAFARI developer docs, October 2026). Connection snippets for each tool are published in the developer documentation.
Which countries does Casafari MCP cover?
The property graph powering Area Insights and Comparables & Valuation covers residential and commercial real estate across 16 countries (US: entering now, selected states, commercial real estate only, via Claro AI) (CASAFARI developer docs, October 2026). You can check full regional details in our help guide on which countries CASAFARI data covers. Two servers have targeted coverage: AgentGraph covers real estate agencies in Germany, while Claro covers commercial premises in Spain and Germany, as well as commercial plots in Germany.
Do I need an API key to use Casafari MCP?
No. Authentication is handled seamlessly via browser-based OAuth 2.1 sign-in when you enter a server URL into your client. The application receives a secure access token directly, removing the need to copy static API keys into config files. If you do not have an existing account, one is established automatically on the Free plan during your first login.
Can an AI assistant change or delete anything in my CASAFARI account through MCP?
Only the watchlist. The data tools in Casafari MCP are read-only; only AgentGraph's watchlist tools save or remove entries in your organisation's shared watchlist. AI assistants can inspect market trends, search comparable listings, and gather agency insights, but cannot alter, add, or erase account settings, saved listings, or valuation records. You retain full control and can revoke server access at any time on platform.casafari.com.
What is the difference between Casafari MCP and the CASAFARI Property Data API?
Both options provide access to the same core European property data, but cater to different interfaces. The REST API at api.casafari.com is engineered for custom software applications using conventional endpoints and static API keys. Casafari MCP formats that same data graph into tools tailored for LLM reasoning inside AI assistants, using OAuth authentication and token-based usage. The complete mapping between REST endpoints and MCP tools is available in the developer tool catalogue.
Try Casafari MCP with 1,000 free tokens
Sign in at platform.casafari.com, add a server URL to your assistant and ask your first question. No card needed.