Bridge Town

official

Bridge Town is an MCP-native, git-versioned financial modeling platform for FP&A teams and finance leaders. AI agents use Bridge Town tools to create projects, write Python model files, run models in isolated cloud sandboxes, query data, write outputs to Google Sheets, create dashboards, branch scenarios, and collaborate with teammates.

GitHub

What can you do with Bridge Town MCP?

Create and manage financial modeling projects — Use search_tools to discover project tools, then create projects, write Python model files, and run them in isolated cloud sandboxes.
Write model outputs to Google Sheets — Push results from model runs directly into Google Sheets that the user has authorized through Bridge Town's scoped access.
Branch scenarios and collaborate — Create branches for alternative assumptions, open pull requests, and share work with teammates using version-controlled models.
Discover tools and skills for FP&A tasks — Use search_tools and get_tool to navigate the 75-tool surface, or retrieve skill templates via get_skill and skills:// resources.

Documentation

Bridge Town MCP Connector

Build financial models as code. Cloud execution, GSheets MCP, version control, collaboration.

This repository contains the public documentation and registry metadata for the hosted Bridge Town MCP connector. The production server implementation is proprietary and is operated by Bridge Town.

Website: https://bridgetown.builders
App: https://app.bridgetown.builders
Documentation: https://docs.bridgetown.builders
MCP endpoint: https://api.bridgetown.builders/mcp
Transport: MCP Streamable HTTP
Authentication: OAuth for hosted connector flows; bearer tokens for clients that need manual configuration

What Bridge Town Does

Claude.ai is the most common connection path and uses OAuth. Claude Code, Claude Desktop, Codex, OpenCode, Cursor, and custom MCP clients can connect to the same Streamable HTTP endpoint with a Bridge Town API token.

Bridge Town does not run a server-side LLM or proxy your prompts. Your selected AI client and model provider run the conversation; Bridge Town receives structured MCP tool calls.

Connect

Claude.ai

Open Claude.ai.
Go to Customize -> + -> Add custom connector.
Paste:
```
https://api.bridgetown.builders/mcp
```
Name the connector Bridge Town.
Approve the Bridge Town OAuth prompt.

No API token is required for Claude.ai.

Generic MCP Client

For clients that support Streamable HTTP plus headers:

{
  "mcpServers": {
    "bridge-town": {
      "transport": {
        "type": "http",
        "url": "https://api.bridgetown.builders/mcp",
        "headers": {
          "Authorization": "Bearer btk_YOUR_TOKEN"
        }
      }
    }
  }
}

Generate tokens from https://app.bridgetown.builders/connect. Tokens start with btk_ and can be revoked from the same page.

For clients that only support spawn-based MCP configuration, use mcp-remote:

{
  "mcpServers": {
    "bridge-town": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://api.bridgetown.builders/mcp",
        "--header",
        "Authorization: Bearer btk_YOUR_TOKEN"
      ]
    }
  }
}

See Getting Started for fuller setup instructions, Google Sheets workflows, skills, and troubleshooting.

MCP Surface

Bridge Town exposes a production MCP surface for financial modeling workflows:

75 tools across projects, models, runs, data, Google Sheets, dashboards, branches, pull requests, sharing, templates, billing, and metadata discovery
MCP resources including health://status, skills://, skill://{name}, templates://, and template://{name}
Workflow prompts and skill templates for common FP&A tasks

For large tool surfaces, start with these discovery tools:

search_tools - find the right Bridge Town tool or skill for a task
get_tool - retrieve a full tool schema
get_skill - compatibility bridge for clients that cannot read MCP resources

Registry Metadata

The canonical registry manifest for this public listing is server.json. It points to the hosted remote MCP server:

https://api.bridgetown.builders/mcp

This repo exists so MCP directories that require a GitHub URL can link to a stable public source of documentation and metadata without requiring Bridge Town to open source its commercial server.

Local Stdio Shim

Most hosted clients should connect directly to the Streamable HTTP endpoint. For registries or local MCP hosts that require a stdio binary, this repository also provides a minimal shim:

BRIDGE_TOWN_API_TOKEN=btk_YOUR_TOKEN npx github:Bridge-Town/financial-modeling-mcp

The shim exposes bridge-town-mcp over stdio and forwards MCP JSON-RPC messages to:

https://api.bridgetown.builders/mcp

Configuration:

Variable	Required	Description
`BRIDGE_TOWN_API_TOKEN`	yes	Bridge Town API token from https://app.bridgetown.builders/connect
`BRIDGE_TOWN_MCP_URL`	no	Override the hosted MCP URL, mainly for local development

The shim does not read or write local files and does not spawn child processes. It only makes HTTPS requests to the Bridge Town MCP API.

Security

Bridge Town is tenant-isolated, OAuth-backed, and designed for financial data. Financial model execution runs in isolated containers with no outbound network. Google Sheets access uses Google's drive.file scope so Bridge Town can only access Sheets a user selects through the Bridge Town flow or Sheets created by Bridge Town.

Report security issues privately. Do not open public GitHub issues for vulnerabilities. See SECURITY.md.