FeasibleHQ is where you build the financial model and write the feasibility study in the same place. Every number in the text is a live reference to the model, so when an assumption changes, the study changes with it. There is nothing left to re-check by hand.
That conclusion paragraph is a real section of a study. Drag the tariff.
We are not open to users yet. Write to us and we will tell you the day it is.
At a tariff of 61.40 €/MWh and the P50 yield, the project produces a net present value of — at a 7.93% cost of capital, with a project return of —.
Minimum debt‑service coverage is — against a 1.20× covenant, in —. The project services its debt in every year of the tenor.
Vale do Sol is an illustrative project and the model behind this paragraph is a simplified one, computed in this page. The engine inside the product is a different order of thing — but the property being shown is the real one: the prose holds references, never numbers.
A financial model in a spreadsheet, a document in a word processor, and a folder of source material meant to justify both.
They drift from the first day. An assumption is corrected in the spreadsheet and the sentence in section four still quotes the old number.
Nobody notices until a lender’s analyst does. By then the three of them disagree, and nobody can say which one the study actually stands on.
The model is computed by a deterministic engine with its own test suite, and the document’s figures are references to that model rather than numbers typed into prose.
The documents you upload are read, quoted and turned into assumptions you approve one at a time. The finished study can then be checked by whoever receives it, in their own browser, without contacting you.
The methodology available today is renewable energy generation, covering eleven standard sections from the executive summary to the recommendation. A study can add sections of its own, and methodologies are written as data rather than as code.
Each one is a place where a study built in a spreadsheet and a word processor costs a day. Under each is the part of FeasibleHQ that takes the day back.
Three weeks after you sent it, the bank wants to see the study at a lower tariff and a fifteen-year tenor instead of eighteen. In a spreadsheet-and-Word study that is a day of reading every section again to find which sentences quoted the numbers you have just changed.
A term sheet, a yield report, an EPC quotation and a grid connection offer. Somewhere inside them are the twenty numbers your model needs, and in six weeks somebody will ask which page each one came from.
You know what belongs in the technology section. Writing it out again, for the fourth project this year, in the register a credit committee reads, is the part that takes the afternoon.
A study that shows only the base case gets that question in the first meeting. Answering it honestly needs a second model, and a second model is stale within a week of the first one changing.
Whoever receives the study will rebuild the model in their own spreadsheet, and will find any figure in the text that no longer matches it. That conversation decides whether they trust the rest of the document.
Same technology, a different site and sponsor. Everything you settled the first time, the structure, the voice, the way you present coverage to a credit committee, is worth having again without copying a file and hunting through it for last quarter’s numbers.
Six parts, built to work as one thing rather than as six features that happen to ship together.
Enter capacity, yield, tariff, capital and operating costs, debt, equity, tax and the macro assumptions. Everything downstream is computed and recomputed as you type.
Upload the term sheet, the yield report, the EPC quotation, the tender pack, the spreadsheet the developer sent. PDF, Word, Excel, CSV and plain text.
“… a fixed tariff of EUR 61.40 per MWh applies for the first fifteen years from the commercial operation date …”
PPA_term_sheet.pdf · page 4An assistant that drafts each section against the methodology, your brief, your own uploaded documents and the way your practice writes. Every figure it produces is a reference to the model, never a typed value.
The project returns a net present value of 10.28 € m against a cost of capital of 7.93%. Coverage holds above the covenant throughout, at a minimum of 1.27× in year 1, and improves thereafter as the senior debt amortises.
The methodology gives you the standard sections and tells you what a reader expects to find in each one. You write; the figures come from the model.
A study that shows only the base case gets asked what happens at P90 before anyone has finished reading it. Here a case is your model with a few assumptions moved, so it cannot drift from the study around it.
| Base | P90 | Capex +10% | |
|---|---|---|---|
| Net present value | 10.28 | 5.49 | 7.42 |
| Project IRR | 11.1% | 9.6% | 10.1% |
| Minimum DSCR | 1.27× | 1.14× | 1.27× |
When the study is finished the figures resolve once, at the instant the files are produced, from the model as it stands.
A tool you use once is a demo. These are the things that pay off from the second study onward.
One search box across every study your practice has produced, matching the words rather than the file names.
Save a paragraph you are happy with and reuse it in the next study. Its figures stay bound, so the previous client’s numbers cannot travel into the new document.
Duplicate a finished project, change the client and the site, and keep the structure, the voice and the model shape you already settled.
Two standing notes — your own voice and the firm’s house conventions — go into every draft, so studies from one office read like one office.
Which sections are written, which figures are still uncited, and the single next step — so a study in progress never has to be reconstructed from memory.
A risk matrix with an owner column, a permit schedule with dates, the site layout, and charts drawn from the model — all of them printed in the finished study.
Studies and uploaded documents belong to your organisation and are readable only by the people in it. Colleagues join by invitation, and nothing is shared outside the practice by default.
The whole workspace in both, because a study gets written at ten in the morning and at eleven at night. The choice is remembered between visits.
Every one of these can be dropped into a sentence, a table cell or a downside case, and every one of them stays tied to the model that produced it.
Valuation, coverage, cost of capital, the three statements, break-even and levelised cost — from the same engine a recipient re-runs when they verify the document.
Six steps, in the order a consultant actually works.
What is being built, where, who buys the output, what the lender has already said. The answers go into the model, not into a side document that goes stale.
Term sheets, yield reports, quotations, tender documents. The assistant reads them and proposes assumptions for you to accept or reject, each with the sentence it came from.
Capacity and yield, capital cost, operating cost, debt, tax, macro. One model, rechecked as you type, with the coverage a lender will ask about visible while you work.
P90 yield, a cost overrun, both together. Each case is the same model with a few assumptions moved, and its figures can be quoted in the text like any other.
The assistant drafts each section against the methodology and your own documents. You keep, edit or discard. Tables and drawings sit in the document where they belong.
PDF, Word and Excel, produced together from one model. The seal is taken over the model that produced the figures, so the document can prove itself.
Not adjectives. Each of these is verifiable by the person holding the document.
The sentence stores {{valuation.npv}}. It has never held a copy of the value, so it cannot be holding an old one. Correct the assumption and the sentence is already correct.
The arithmetic is deterministic and tested. The assistant is a writer with reading rights, not a calculator you have to trust.
Your reader does not have to take the figures on faith, or ask you for the model. They reproduce them from the file you sent, in their own browser, with no account and nothing uploaded.
Drop in a spreadsheet you already have. Fourteen structural checks a credit committee runs before it reads a word of the story: balance‑sheet integrity, cash‑flow reconciliation, the basis the coverage ratio is calculated on, circular references, hardcoded values sitting inside formulas. It runs entirely inside your browser — the page is served with a policy that forbids it from making any network request at all, so nothing can be uploaded even by mistake. Useful on a model somebody else built, before your name goes anywhere near it.
A work platform for producing feasibility studies. It holds the three things that normally live in three separate files — the financial model, the document, and the evidence behind them — and keeps them in step with each other.
No. Every figure comes from a deterministic financial engine with its own test suite. The assistant writes prose and cites figures from the model; it is not permitted to write a number, and if its output contains one you are shown it rather than it being quietly accepted.
It drafts each section against the methodology, your brief, the documents you uploaded and the way your practice writes, and you keep, edit or discard what comes back. It appends rather than overwriting, so your own sentences survive. Every figure it produces is a reference to the model; its output is inspected for anything that looks like a typed number or a reference to a figure that does not exist, and both are reported to you.
Every sentence that cites it changes with it. A figure in the text is a reference to a place in the model, not a number that was typed there, so correcting the yield corrects the executive summary, the risk section and the conclusion at the same time. Tables and downside cases move too.
Renewable energy generation, with a full methodology covering eleven standard sections from the executive summary to the recommendation. You can also add sections of your own to any study. Further methodologies are written as data rather than as code, so they arrive without a rebuild.
Yes. Upload term sheets, yield reports, EPC quotations, tender documents and spreadsheets — PDF, Word, Excel, CSV and plain text. The assistant reads them and proposes model assumptions, each one carrying the exact sentence it came from, checked against the file before you see it. You accept or reject each proposal; nothing is written into your model on its own.
A typeset PDF for the client, an editable Word document with real Word tables, and the model itself as an Excel workbook with live formulas. All three are produced from one model at one moment, so they cannot tell different stories.
A sealed study carries a SHA-256 seal over its assumptions, its parameters and every figure computed from them, and the PDF carries the machine-readable model inside it as an attachment. The recipient drops the file on the verification page, which re-runs the model in their own browser and reproduces every figure in the document, digit for digit. They never have to contact you, and nothing is uploaded anywhere.
They are stored privately and are readable only by your own organisation. When you ask the assistant to draft a section or to read a document, the passages it needs are sent to the model provider for that request; nothing is sent at any other time. The free spreadsheet audit sends nothing at all — that page is served with a policy that forbids it from making any network request.
That is the point of it. Start a new study from an old one, keep passages you are happy with and reuse them — with their figures still bound, so the previous client’s numbers cannot travel into the new document — and search across everything your practice has written.
It is free while we are in early access. There is no card and no billing to cancel.
Free while we are in early access. Bring a project you are already working on, or the documents for one, and see how far it gets in an afternoon.