Sample plans
Whole plans, not excerpts.
Four complete plans, readable end to end. Each one was built during this site's build from a real set of assumptions, run through the same engine your plan uses, and written by the same generator.
Three of these four would not export. We are showing you that rather than hiding it, because the reason is the product working.
For a bank or SBA lender
Chapel Street Kitchen
Providence, Rhode Island
A 78-cover neighbourhood restaurant seeking SBA 504 financing against a kitchen build-out. The plan a credit committee reads first.
- Year 3 revenue
- $2M
- Year 3 net margin
- 7.0%
- Coverage
- 4.14×
- Readiness
- 59/100Not ready to send
For investors
Meridian Supply
Portland, Oregon
A direct-to-consumer brand raising a first institutional round. Unit economics and a market section built from the buyers up.
- Year 3 revenue
- $1.7M
- Year 3 net margin
- -4.3%
- Break-even
- Month 34
- Readiness
- 59/100Not ready to send
For a bank or SBA lender
Harborlight Early Years
Duluth, Minnesota
A licensed childcare centre with ratio-mandated staffing. What a plan looks like when the cost base is set by regulation.
- Year 3 revenue
- $1.2M
- Year 3 net margin
- 7.5%
- Coverage
- 2.40×
- Readiness
- 59/100Not ready to send
For internal planning
Atlas & Vance
Chicago, Illinois
A professional-services firm planning internally, with no debt and no raise. Utilisation is the whole business.
- Year 3 revenue
- $1.2M
- Year 3 net margin
- 9.5%
- Break-even
- Month 1
- Readiness
- 94/100Ready to send
Why three of them are blocked
We did not invent any competitors.
A sample plan is the easiest place in the world to write three plausible competitors with three plausible prices. Nobody would check. It would also contradict the one guarantee this product is built on, so the competitive section of every sample is empty.
The consequence is visible above: the review blocks the three plans whose reader expects competitive evidence, and they score 59. The fourth is an internal plan, where that evidence is not what the reader is testing, so it scores 94 and would export. Same document, same engine, different bar — which is the point.
In a real plan that section is filled from live research, and every row carries a link and the date the price was observed. Undated evidence is excluded rather than quietly accepted.
What is real in these
- The financial model. Sixty months, three linked statements, and a balance sheet that ties in every period.
- The prose. Composed by the same deterministic generator that runs whenever no API key is present — it writes real sentences and invents nothing.
- The review. Scored against the rubric for that plan's specific reader, with the findings shown.
- The document itself. Rendered by the same component that renders a plan you share with an investor.
The businesses are invented, and so are the population figures behind each market build — they are stated as the owner’s own assumptions, which is what they would be in a first draft. With a live key the prose is written by Claude instead, against the same figures.
Yours will not be invented.
The intake asks about your business and the engine builds the model from your answers. Free to generate and read.