Platform
Get started

Casafari is the AI agent-native real estate data intelligence platform. The most complete property index in Europe: a deduplicated, cleaned property graph of residential and commercial property, for sale and for rent, in 16 countries.The most complete property index in Europe: a deduplicated, cleaned property graph.

How the graph is built

Agent recruiting shortlist

Find the agents and agencies worth hiring or partnering with.

You get: Shortlist with profiles, watchlist. It uses AgentGraph AI, takes 160 tokens a run (€6.40 at €0.04 a token) and runs monthly. For: Estate agency or network.

Status: runs today.

What it does

#StepProductTokens
1agentgraph_find_locationAgentGraph AI0
2agentgraph_screen_agentsAgentGraph AI125
3agentgraph_agent_profile × 5AgentGraph AI25
4agentgraph_save_to_watchlist × 5AgentGraph AI10

One run: 160 tokens, €6.40. Each call costs the same over MCP and REST, so the run always costs what the table says. How many tokens your volume needs, and plans and credits.

Where it runs

Germany. Ask for another country and the plan names what is missing.

Run in Claude

This playbook calls one Casafari MCP server. Add it to Claude, ChatGPT or any MCP client, and sign in the first time you use it (a Free account, no card, is created then; console).

  • AgentGraph AI: https://platform.casafari.com/mcp/agentgraph

In Claude Code:

bash
claude mcp add --transport http casafari-agentgraph https://platform.casafari.com/mcp/agentgraph

Then ask:

text
Run the agent recruiting shortlist playbook for Germany every month: call agentgraph_find_location, then agentgraph_screen_agents, then agentgraph_agent_profile, then agentgraph_save_to_watchlist, in that order, and tell me what each step cost in tokens.

Get the API calls

The same steps as REST calls, in order. Create a platform API key in the console and export it as CASAFARI_API_KEY; each call takes the parameters its operation documents in the REST reference.

bash
# 1. AgentGraph AI, 0 tokens a call
# MCP only for now: call agentgraph_find_location in your assistant.

# 2. AgentGraph AI, 125 tokens a call
# MCP only for now: call agentgraph_screen_agents in your assistant.

# 3. AgentGraph AI (5 times), 5 tokens a call
# MCP only for now: call agentgraph_agent_profile in your assistant.

# 4. AgentGraph AI (5 times), 2 tokens a call
# MCP only for now: call agentgraph_save_to_watchlist in your assistant.

Other playbooks

Tip: add .md to any URL to read it as Markdown.