Connect your AI tools to Pi to turn your natural language conversations into location intelligence insights. The Placer MCP enables your assistant to run full Placer analyses--foot traffic, trade areas, visitor demographics and journeys, cross shopping, spend--and bring the answer back into your conversation with a link to the full analysis in the Pi platform. Ask Pi directly from Claude, ChatGPT, and other AI assistants that speaks the Model Context Protocol to speed your location intelligence analysis within your own workflows.

Placer MCP is an enterprise feature and must be enabled per account to access. See Requirements section below.

Put Pi To Work For You

Pi answers business questions across multiple audiences:

Pi is for location intelligence only. General web lookups, company facts, coding, math, and writing are outside its scope — your assistant should handle those itself.

Requirements

• A Placer.ai account with Pi enabled.
• The MCP access entitlement on your account. This is granted per account and is not included by default — reach out to your Placer representative to discuss.
• An AI assistant that supports remote MCP servers over streamable HTTP with OAuth (Claude, ChatGPT, GitHub Copilot, Cursor, and others).
• Connection with your normal Placer credentials. Pi sees exactly the data your Placer user is entitled to.

Connecting Your Assistant

Server URL: https://pi.placer.ai/mcp
Transport: Streamable HTTP
Authentication: OAuth 2.0 (Placer SSO)

Every host follows the same process: add the server URL, then complete the Placer sign-in that opens in your browser.

Claude (web and desktop)

1. Open Settings → Connectors → Add custom connector.
2. Name it Placer and paste the server URL.
3. Click Connect and sign in with your Placer account.
4. In a new chat, confirm Placer appears in the tools menu.

Claude Code

claude mcp add --transport http placer {{pi.placer.ai/mcp}}
Then run /mcp in Claude Code and choose Authenticate to complete the Placer sign-in.

ChatGPT

1. Enable Settings → Connectors → Advanced → Developer mode.
2. Create a connector, paste the server URL, choose OAuth authentication.
3. Click Connect and sign in with your Placer account.

VS Code / GitHub Copilot

Add to .vscode/mcp.json (or run MCP: Add Server from the Command Palette):

{"servers": { "placer": { "type": "http", "url": "{{pi.placer.ai/mcp}}" } }}
Start the server from the editor and complete the Placer sign-in when prompted.

How an analysis runs

Pi does extensive analytical work: it finds the right places, pulls the data, checks it, and writes the answer. A full analysis normally takes several minutes. Because of that, the connector is asynchronous:

1. Your assistant calls ask_pi with your question. It returns immediately with a handle — conversation_id, message_id — and status: "working".
2. Your assistant calls get_pi_result repeatedly. While Pi works, each call returns status: "working" plus the latest progress steps, so your assistant can tell you what Pi is doing.
3. When Pi finishes, get_pi_result returns status: "done" with the written answer, any data-honesty notes, web sources if used, and a deep link into Pi.

Just ask your question to your assistant in natural language. The above flow is what your assistant does under the hood.

Follow-ups. Pass the previous conversation_id back to Pi and Pi reuses the data it has already pulled, so the numbers stay consistent and the follow-up is faster. Write follow-up questions in full — Pi does not remember the earlier wording.

Tools

Pi has two read-only tools. Pi analyzes and reports; the connector never creates, modifies, or deletes anything in your Placer account.

ask_pi — Ask Pi

Starts an analysis and returns a handle to poll.

Returns: status (working), conversation_id, message_id, poll_after_seconds, pi_deep_link.

get_pi_result — Get Pi Result

Collects the answer from an analysis started by ask_pi.

Returns while running: status (working), progress, progress_cursor, poll_after_seconds, pi_deep_link.

Returns when finished: status (done), narrative (Pi's written answer), analysis_notes (data-honesty caveats), sources (web sources cited, if any), pi_deep_link, conversation_id.

On failure: status (error) and a human-readable error message.

Limits

Analyses consume Pi credits from your account's balance, the same as analyses run in the Pi app.

Data and privacy

• Authentication is OAuth 2.0 against Placer SSO. The connector never asks for or stores your password, and access is revoked when your Placer account is.
• Authorization is enforced on every single tool call: your Placer identity, your account's entitlements, and your data permissions. If access is revoked, the next call is denied immediately.
• Read-only. Neither tool writes to Placer data.
• Conversations you start through the connector appear in your Pi history, so you can reopen any analysis in the Pi app via its deep link.
• Handling of personal data is governed by the Placer.ai Privacy Policy.

Troubleshooting

Support

Reach out to your Placer representative (best), or use the Placer.ai contact form.