Factori MCP works with any MCP-compatible AI client.

Pick your client below and follow the steps, setup takes under two minutes and uses OAuth, so no API keys are required.

MCP server URL https://mcp.factori.ai/mcp

Claude.ai/ Desktop

Open the Claude desktop app, or go to claude.ai and sign in
Click your profile (bottom-left), then Settings
Open the Connectors tab
Click Add custom connector In Name type Factori. In URL paste https://mcp.factori.ai/mcp
Click Add, then Connect Your browser opens console.factori.ai — sign in and click Approve.

You are redirected back to Claude automatically Start a new chat and look for the tools icon — Factori is ready

Team / Enterprise: an Owner must first add it under Organization settings → Connectors. To verify, start a new chat and ask "How many credits do I have?" — Claude should return your balance.

VS Code

Open your project in VS Code
In the file sidebar, create a .vscode folder if it doesn't exist, then create a file inside it called mcp.json
Paste the following and save:

{ "servers": { "Factori": { "type": "http", "url": "https://mcp.factori.ai/mcp" } } }

Open the Chat view and switch to Agent mode
On first use, VS Code prompts you to sign in — your browser opens console.factori.ai; Approve and you are sent back

Note: verify the root key is "servers" (not "mcpServers") if the config does not load.

ChatGPT

Requires a Plus, Pro, Business, or Enterprise plan.

Click your profile → Settings → Connectors
If you don't see a custom option, open Advanced and turn on Developer mode
Click Add custom connector
Name it Factori and paste https://mcp.factori.ai/mcp
Click Create / Connect
Your browser opens console.factori.ai — sign in and Approve
In a chat, open the connectors menu and switch Factori on

OpenAI Codex

Open your terminal and run: codex mcp add Factori --url https://mcp.factori.ai/mcp
Sign in by running: codex mcp login Factori
Your browser opens console.factori.ai — sign in and Approve
Restart Codex. Factori's tools are now available

Gemini CLI

Open the settings file in any text editor (create it if it doesn't exist):
macOS / Linux: ~/.gemini/settings.json
Windows: %USERPROFILE%.gemini\settings.json
Paste the following and save:

{ "servers": { "Factori": { "type": "http", "url": "https://mcp.factori.ai/mcp" } } }

Restart Gemini CLI
Run /mcp auth Factori
Your browser opens console.factori.ai — sign in and Approve
Run /mcp — a green Factori — Ready confirms the connection

Note: the consumer Gemini app does not support custom servers yet. Use Gemini CLI, Antigravity or Gemini Enterprise.

Cursor

Press Cmd+, (Mac) or Ctrl+, (Windows) to open Settings
Go to Tools & MCP and click New MCP Server — this opens mcp.json
Paste the following and save:

{ "servers": { "Factori": { "type": "http", "url": "https://mcp.factori.ai/mcp" } } }

Sign in when prompted — your browser opens console.factori.ai; Approve and you are sent back
A green dot next to Factori in Tools & MCP confirms the connection

Note: If tools do not appear after editing mcp.json, fully quit (Cmd+Q on Mac) and reopen Cursor — hot-reload is unreliable for config changes.

Windsurf

Press Cmd/Ctrl+, → Cascade → MCP Servers → View raw config, or open the file directly:
macOS: ~/.codeium/windsurf/mcp_config.json
Windows: %USERPROFILE%.codeium\windsurf\mcp_config.json
Paste the following and save:

{ "servers": { "Factori": { "type": "http", "url": "https://mcp.factori.ai/mcp" } } }

Fully quit and reopen Windsurf, closing the window alone will not reload it.
On first use, your browser opens console.factori.ai, sign in and Approve.

Antigravity (Google)

In the Agent panel, click the ... menu → MCP Servers → Manage MCP Servers
Click View raw config — this opens mcp_config.json
Paste the following and save:

{ "servers": { "Factori": { "type": "http", "url": "https://mcp.factori.ai/mcp" } } }

Return to Manage MCP Servers and click Refresh
On first use, your browser opens console.factori.ai — sign in and Approve

Verify your setup

Once connected, you do not need any special commands or syntax. Just ask your AI assistant a question in plain English. If Factori is live, it will call the relevant data tools and return a structured answer within seconds.

To confirm everything is working, try one of the prompts below.

Role	Prompt to try	What a working response looks like
Retail / Site Selection	"Which neighborhoods in my metro show rising foot traffic but low competitor saturation?"	A ranked list of neighborhoods with foot traffic trend data and competitor density scores
QSR / Franchisee	"How has foot traffic at my locations trended over the last 12 weeks vs. category averages?	A week-by-week trend comparison across your locations against the category baseline
Media Planner	"Show me the top 10 US metros ranked by young professional concentration and peak mobile hours"	A metro ranking with demographic breakdown and daily mobility peak windows
Real Estate	"Which neighborhoods show consistent traffic growth but low occupancy rates?	A shortlist of neighborhoods with traffic trend, occupancy gap, and opportunity score
Any role	"How many credits do I have?"	Your current credit balance, the simplest confirmation that the connection is live

If your AI returns a structured, data-backed answer to any of these, your setup is complete.

Troubleshooting

If something is not working, find your issue below.

Issue	Fix
Connector won't authorize	Make sure you are signed in to Factori in another browser tab, then retry the OAuth flow
"Insufficient credits" error	Ask your AI: "How many credits do I have left?" to check your current balance
Tools don't appear in the client	Restart your client and verify the connector is active. In VS Code, confirm you are in Agent mode
VS Code: config not loading	Verify the root key in mcp.json is "servers" not "mcpServers". Run MCP: List Servers from the Command Palette to debug
Cursor / Antigravity: npx command not found	Install Node.js 18+ from nodejs.org
Cursor: tools not appearing after editing mcp.json	Fully quit (Cmd+Q on Mac) and reopen. Hot-reload is unreliable for MCP config changes
Bridge clients: mcp-remote keeps re-prompting for login	Delete cached tokens by running: rm -rf ~/.mcp-auth/ — then retry
Gemini CLI: server not connecting	Verify you used "httpUrl" not "url" in settings.json. Run /mcp reload to retry
Antigravity: tools not showing after config edit	Go to Manage MCP Servers and click Refresh. If that fails, uninstall and re-add the entry

Disconnect Factori

You can remove the connector at any time. Your Factori account and data are not affected.

Client	How to disconnect
Claude.ai / Desktop	Settings → Connectors → Factori → Uninstall
Claude Code	Run: claude mcp remove factori
VS Code	Command Palette → MCP: List Servers → remove the entry
Gemini CLI	Remove the "factori" block from ~/.gemini/settings.json
Cursor / Antigravity / others	Remove the "factori" block from your MCP config file and restart the client

You can reconnect at any time by repeating the setup steps for your client.