AI models sound certain about Canadian real estate, even when they're guessing at the catchment, the permit history or the flood zone. Neighbourly MCP connects any AI application to our location data over the Model Context Protocol, so those answers come from a sourced record.
Neighbourly MCPA general model can describe Canada beautifully, but it can't look up a catchment boundary, a permit file or a lake polygon unless it has a connection to them. Without one, a wrong answer reads exactly like a right one.
| The question | A model from memory | With Neighbourly MCP |
|---|---|---|
| Which school catchment is this address in? | Names a nearby school. Proximity and catchment are not the same thing. | The catchment the address falls inside, from school board boundaries. |
| Have there been permits on this home? | Has no way to know, or describes what is typical for the area. | The permit records on file, with type, status, date and value, where the municipality is covered. |
| Is this cottage on a lake? How deep is it? | Repeats what it can find about the lake, which may be a different lake with the same name. | The lake polygon the coordinate matches, with depth, thermal regime and stocking records. |
| Is this neighbourhood flood-prone? | Offers a general impression of the region. | Flood zone and hazard overlays for that location, ready to cite. |
| What is this neighbourhood called? | Mixes up real estate names, wards and postal areas. | A stable neighbourhood ID within a consistent Canadian boundary hierarchy. |
The model reads each figure from a Neighbourly record, so your team can trace a claim back to where it came from.
When a record isn't there, the tool says so. Your assistant can tell the user what it doesn't know instead of filling the gap with something plausible.
Skip building and maintaining your own Canadian location pipeline. Connect once, and every data layer on your plan is available to your AI features.
MCP is an open standard for connecting AI applications to outside tools. We expose Neighbourly data as read-only tools that the model calls in the middle of a response, so it reads the figure instead of recalling it. You keep your model, your prompts and your product.
GET/v1/address?q=1184+Hemlock+Point+RoadGET/v1/environment/lakes/find?lat=45.0413&lng=-78.5128GET/v1/fishing_access_pointGET/v1/permits?address=1184+Hemlock+Point+RoadIllustrative exchange. Exact tool names and parameters are shared with your access details.
Your AI gets the same data as our API, built for Canada from the ground up: Statistics Canada geographies, Canadian address formats and bilingual names.
Which neighbourhood, ward and census area is this in?
Is this a real Canadian address? Where exactly is it?
What is listed here, and what are the property details?
What's this neighbourhood like, and how is it defined?
Who lives in this area? Income, age, household size?
What work has been permitted at this address?
Which lake is this? Crown land, trails, flood or wildfire exposure?
What does this home cost to run compared with others nearby?
Which catchment is this, and which schools are close?
What businesses and services are in this trade area?
Availability varies by layer and region. Energy data covers Ontario, with a pilot in BC. See coverage for details.
Give agents a local expert that answers client questions about schools, permits and neighbourhoods with real records.
Generate listing copy from checked location facts, so a wrong catchment or lake never reaches the page.
Let buyers ask for "a quiet lake with deep water under two hours from Toronto" and answer from boundary and environmental data.
Offer members grounded AI features without each brokerage sourcing its own location data.
Bring sourced hazard, boundary and neighbourhood context into underwriting and review workflows, with people making the decision.
Let analysts screen trade areas in plain language, with demographics and business data behind every answer.
Location data informs real decisions about homes, money and communities. The connection is built to match that responsibility.
The tools look data up. They cannot create, change or delete anything.
Your key opens the data layers you already license, and nothing else.
Client data is never used to train AI models shared with other clients, as set out in our Responsible AI commitments.
Neighbourly data is hosted in Canada, on the same infrastructure as our API.
Demographics are aggregated to census geographies. People remain accountable for decisions, and AI supports their judgment.
A connection that gives AI applications read-only tools for looking up Neighbourly location data over the open Model Context Protocol. The model calls a tool when it needs a fact, and answers from the record it gets back.
Any client that supports MCP, such as Claude, Cursor and VS Code, and applications built on the OpenAI or Anthropic APIs.
Data from the ten Neighbourly layers: boundaries, addresses, real estate, neighbourhoods, demographics, building permits, environment, energy, schools and businesses. Access follows your plan, and availability varies by layer and region.
Neighbourly covers all provinces and territories, with availability varying by layer. Energy data covers Ontario, with a pilot in BC, and permit coverage depends on the municipality. The coverage page lists the details.
The tool returns no result, so your AI can say the information is unavailable rather than inventing it.
Web search returns pages for a model to interpret. Neighbourly MCP returns structured records matched to a coordinate or address, including data like catchment boundaries, permit records and lake geometry that a page search may never surface accurately.
No. The tools are read-only and cannot create, change or delete anything. Client data is never used to train AI models shared with other clients.
Contact our team. We'll review your use case and send you the connection details and key.
Tell us what you're building, and we'll get you connected.