What goes into the study
Every input and formula is the same one a professional feasibility study runs, just organized into a sequence instead of twenty spreadsheet tabs.
Fixed investment
What it takes to open the doors, itemized and totaled.
Operating costs
Payroll, rent, stock and energy, split into fixed and variable.
Returns
Break-even, payback, ROI, NPV and IRR, recomputed as you type.
Starting points for cafés, restaurants, shops, software startups, service firms and manufacturing, or start blank.
From numbers to a plan
Once the study is confirmed, describe the project in your own words. Foundation drafts a business plan around it (market, competition, operations, financials) and saves it to the project as a Word file you can download at any time. Every figure in it comes from the study you built, not from the AI.
Using Claude, ChatGPT or another assistant with a frontier model? It can write a fuller plan from the same numbers. See how
Work through the AI you already use
Foundation is also an MCP server. Point your own agent at it, sign in once, and it can read or edit the same study directly. No copying numbers back and forth between a chat window and a browser tab.
Five open-source skills teach it the analyst’s method: feasibility, market research, pricing, the business plan and the pitch deck. They also work without the server, and hand back a project file you can import here.
$ claude mcp add --transport http foundation https://getfndtn.com/api/mcp$ claude> list my Foundation projectsThe first request opens your browser to sign in. Nothing more to run.