Property, people & company data,
ready for your AI agents.

Create a new Custom GPT

In ChatGPT → Explore GPTs → Create → switch to the Configure tab.

Add a Plotted action

Scroll down to Actions → Create new action → in the Schema box, paste this URL:

https://plotted.to/openapi.json

ChatGPT auto-imports all 24 operations.

Add your API key

Under Authentication → pick API Key → Bearer → paste your plt_live_***.

Suggested system prompt

You are a U.S. property, people, and company research assistant powered by Plotted
(150M+ property records + nationwide professional profiles).

Property & addresses:
- Use /v1/geocode to resolve an address to lat/lon when needed.
- Use /v1/parcels/nearby for radius searches around a point.
- Use /v1/views/absentee, /v1/views/trust, /v1/views/homeowners for curated lead segments.
- Combine filters: state, county, is_trust, has_contact, year_built_min/max, value_min/max, beds_min, baths_min, sqft_min/max.
- For bulk enrichment, use /v1/match (up to 10,000 addresses per call).

Enrichment, people & companies:
- Use /v1/enrich/address (address_id or address+zip) for the people, contacts, and distress at an address.
- Use /v1/person/enrich (email, phone, name+zip, or linkedin URL) for a person + their LinkedIn professional profile.
- Use /v1/person/search (anchor: company, name, or linkedin_url; refine with title, role, industry, region) to find people.
- Use /v1/company/enrich (name or ein) for business registry officers/agent + LinkedIn firmographics.
- Call /v1/openapi any time for the full endpoint catalog.

Always cite parcel row_id when referencing specific properties. Redact phone/email digits in
output unless the user explicitly asks for the unmasked value.

Define the Plotted tool

import google.generativeai as genai

plotted_tool = {
    "function_declarations": [
        {
            "name": "plotted_nearby_parcels",
            "description": "Find parcels within radius_mi of a lat/lon. Filters: is_trust, has_contact, owner, year_built_min/max, value_min/max, beds_min, baths_min, sqft_min/max.",
            "parameters": {
                "type": "object",
                "properties": {
                    "lat": {"type": "number"},
                    "lon": {"type": "number"},
                    "radius_mi": {"type": "number"},
                    "is_trust": {"type": "boolean"},
                    "has_contact": {"type": "boolean"}
                },
                "required": ["lat", "lon", "radius_mi"]
            }
        }
        # ... add the other 15 endpoints (or fetch via /openapi.json)
    ]
}

model = genai.GenerativeModel("gemini-2.0-pro", tools=[plotted_tool])

Handle the function call

import requests, os
PLOTTED_KEY = os.environ["PLOTTED_KEY"]

def call_plotted(name, args):
    endpoint_map = {
        "plotted_nearby_parcels": "/v1/parcels/nearby",
        # ...
    }
    r = requests.get(
        "https://api.plotted.to" + endpoint_map[name],
        params=args,
        headers={"Authorization": f"Bearer {PLOTTED_KEY}"}
    )
    return r.json()

response = model.generate_content(
    "Find trust-owned parcels within 1mi of 26.142, -81.795 with phone numbers"
)
# If response.candidates[0].content.parts[0].function_call exists, call_plotted() and send back

Open Cursor Settings → Features → Model Context Protocol

Or edit ~/.cursor/mcp.json directly. Add this block:

{
  "mcpServers": {
    "plotted": {
      "command": "npx",
      "args": ["-y", "@plotted/mcp-server"],
      "env": { "PLOTTED_API_KEY": "plt_live_***" }
    }
  }
}

Restart Cursor, then ask in chat

Example: "Take the addresses in selected.csv, call Plotted's /v1/match for each, and add the owner_name + phone columns." Cursor handles the rest.

Codex, Antigravity, Windsurf, VS Code & every other MCP client

Same server everywhere — point the tool at @plotted/mcp-server. Config locations per platform:

• Antigravity (IDE & CLI) — ~/.gemini/config/mcp_config.json (same mcpServers JSON as above), or IDE agent panel ··· → Manage MCP Servers → View raw config
• Windsurf — ~/.codeium/windsurf/mcp_config.json (same JSON)
• VS Code (Copilot agent mode) — .vscode/mcp.json under a "servers" key
• Continue, Zed, Cline, Roo Code — same mcpServers JSON
• OpenAI Codex — TOML, shown below

[mcp_servers.plotted]
command = "npx"
args = ["-y", "@plotted/mcp-server"]

[mcp_servers.plotted.env]
PLOTTED_API_KEY = "plt_live_***"

Or one command: codex mcp add plotted --env PLOTTED_API_KEY=plt_live_*** -- npx -y @plotted/mcp-server

OpenAPI 3.1 spec

https://plotted.to/openapi.json

Auto-imports all 24 operations, request/response schemas, error envelopes, and the auth scheme (Bearer token).

LangChain

from langchain_community.tools.openapi.utils.openapi_utils import OpenAPISpec
from langchain_community.agent_toolkits.openapi.toolkit import RequestsToolkit
from langchain_community.utilities.requests import RequestsWrapper

spec = OpenAPISpec.from_url("https://plotted.to/openapi.json")
toolkit = RequestsToolkit(
    requests_wrapper=RequestsWrapper(headers={"Authorization": "Bearer " + KEY}),
    spec=spec,
)
agent.tools.extend(toolkit.get_tools())

LlamaIndex

from llama_index.tools.openapi import OpenAPIToolSpec

spec = OpenAPIToolSpec(
    url="https://plotted.to/openapi.json",
    auth_headers={"Authorization": "Bearer " + KEY},
)
tools = spec.to_tool_list()

Direct curl from any prompt-engineering harness

curl https://api.plotted.to/v1/parcels/nearby \
  -G -d "lat=26.142&lon=-81.795&radius_mi=1&is_trust=true" \
  -H "Authorization: Bearer $KEY"