Product overview
The financial modeling workspace for your AI agent and your team
Bridge Town gives finance teams a shared model workspace where assumptions, formulas, source files, and outputs are explicit. Your agent edits the model directly; Bridge Town keeps the project versioned, reviewable, and auditable.
/ core capabilities
What Bridge Town gives your team
Models
Financial models
Build, inspect, and version financial models as Python projects that your own AI agent can read and update through Bridge Town MCP.
Scenarios
Scenario planning
Run forecast scenarios from versioned assumptions so FP&A teams can compare upside, base, and downside cases without losing the audit trail.
Variance
Variance analysis
Explain plan versus actual movement with versioned model logic, source data, and agent-assisted analysis that finance teams can review.
Audit Log
Audit log for AI-built financial models
Review who changed model logic, which runs produced outputs, and how agent-assisted financial models evolved over time.
What Bridge Town provides
- Project-based storage — each financial model is a project with a full Git-backed version history
- Sandboxed execution — models run in isolated containers; no outbound network, no host filesystem access
- AI agent interface — Claude, Codex, Cursor, or any AI agent can read, write, and run models directly through Bridge Town
- Data connectivity — attach Google Sheets or CSV files; DuckDB runs queries in-process without a data warehouse
- Dashboards — one-call HTML dashboard generation from any model run; share a file or a hosted link
Who it's for
Bridge Town is built for senior FP&A analysts and finance teams who want AI agents to do useful work — not just generate text — in their models. It is the infrastructure that makes AI-generated financial models trustworthy, versioned, and collaborative.
Connect your agent to Bridge Town
Start with a structured workspace, then let your own AI agent build, inspect, and revise the finance model with you.