Calculators & quote tools
Turn your pricing into a number buyers trust
Savings, ROI, tax, capacity or quote calculators with every rule sourced and tested.
OWN PRODUCT · LIVE ↗ TotalTax: 15 countries, every figure sourced
Chosen · Tools & portals
Custom AI Tools, Portals & Agent Endpoints
Calculators, configurators, quote tools and client portals that let buyers qualify themselves, with logic you can audit, and an optional endpoint so AI agents can use your service too.
from €8,900 · $9,800 3–8 weeks
Typical project €9k–25k. Fixed scope, agreed in writing before work starts.
Proof on file
OWN PRODUCT · LIVE ↗~1,100
automated tests
TotalTax, my own live calculator: the build refuses to ship if any figure lacks a dated published source.
observed 2026-10-05 · test suite count
03 Fit
Who this is for, and who it isn't.
Built for you if
- You sell on numbers: pricing, savings, tax, ROI, capacity, lead times.
- Your product has options buyers struggle to picture or combine.
- Clients email you all week for files, documents and status updates.
- You expect AI agents to research, quote or book on your clients’ behalf, and want them to get it right.
Not for you if
- You need a contact form or a booking widget. That’s part of a landing page.
- There’s no real logic behind it. A calculator that always ends in “call us” annoys buyers.
- You want a full SaaS platform with billing and multi-tenant admin from day one. Possible (Filyhost is one), but scoped as its own project.
- The rules change daily and nobody owns them. A tool is only as current as the person maintaining its sources.
04 The problem
Your spreadsheet is your best salesperson. Nobody outside the company can use it.
Most B2B firms have one: the pricing sheet, the savings model, the configurator that lives in a sales engineer’s head. Prospects can’t use it, so they email, wait for a quote, and often buy from whoever answered first.
A public tool changes the order of events. Buyers get their own number in a minute, share the link with their CFO, and arrive at your call already qualified. And as buyers delegate research to AI agents, a tool those agents can call directly answers with your numbers instead of a guess.
The hard part isn’t the interface. It’s logic that is right, traceable and tested, so a number on your website never becomes a liability.
Signs you need this
- Every quote takes a day of back-and-forth.
- Your pricing is too complex for a price list, so you don’t publish one.
- Clients email for files and status updates several times a week.
- Sales engineers spend their time on prospects who were never a fit.
- Competitors have a calculator. Yours is a PDF.
- You don’t know what an AI agent says when asked what you charge.
06 What I build
Four kinds of tools, each one already running.
Configurators
Let buyers build it before they buy it
Visual product builders that check what fits, generate the drawing and send the result to sales.
OWN PRODUCT · LIVE ↗ SitShape: 155 sofa shapes in 8 families
Client portals
Give clients a door instead of an inbox
Files, documents, status and approvals behind a one-time code, with no passwords to manage.
OWN PRODUCT · LIVE ↗ Filyhost: send and receive files with expiring links
Agent endpoints
Let AI agents use your service
An MCP server or WebMCP so assistants can query, quote or book, inside limits you set.
OWN PRODUCT · LIVE ↗ Vowko: 9 MCP tools; agents act, never vouch
07 What you get
Deliverables, tagged by the layer they fix.
Scope depends on the tool, but every build includes the logic, the tests, and a way for buyers and agents to use it.
- Layer: Chosen
The tool itself
Calculator, configurator, quote tool or client portal, designed for a phone first and fast on any device.
- Layer: Engine room
Sourced logic with a test suite
Every rule traceable to a source. Automated tests fail the build if a number breaks or a source goes missing.
- Layer: Chosen
Shareable result links
Inputs live in the link, so a result can be forwarded to a colleague or a CFO and reopened exactly.
- Layer: Chosen
Lead capture
Optional: results by email, or a quote request that arrives with every input attached.
- Layer: Chosen
Stripe
Deposits, licences or subscriptions, where the tool sells directly.
- Layer: Found
MCP server or WebMCP (optional)
Scoped, logged, rate-limited access for AI agents, with actions only a human may confirm.
- Layer: Cited
Method pages
Crawlable pages that explain how the tool calculates, with sources, so assistants can cite the method, not just the number.
- Layer: All layers
Security and privacy
Rate limits, bot protection, Cloudflare Access on admin, and nothing stored that you don’t need.
08 How it runs
Logic first, interface second.
The engine is built and tested against your real examples before anything is designed. If the numbers are wrong, nothing else matters.
Agents at work Human gate: nothing moves on until a person approves
-
Step 1: Week 1
Spec
Inputs, rules, sources and edge cases written down, plus a clickable outline of the screens.
- agent://researcher
- agent://architect
Human gate You approve the spec and the source for every rule.
-
Step 2: Weeks 2–3
Engine
The calculation engine built and tested in isolation, with test cases taken from your real quotes and examples.
- agent://builder
- agent://tester
Human gate You check the numbers against cases you know by heart.
-
Step 3: Weeks 3–5
Interface
Phone-first interface, result pages, sharing and lead capture, on a preview link.
- agent://builder
Human gate You approve the interface on the preview.
-
Step 4: Weeks 5–7
Agents and integrations
Optional MCP endpoint, Stripe, CRM or email integrations. Security, rate-limit and load checks.
- agent://builder
- agent://tester
-
Step 5: Weeks 6–8
Launch and handover
Live on your domain, monitoring on, runbook delivered, your team shown how to update the sources.
- agent://watcher
Human gate You sign off go-live.
09 Proof
Four tools I run myself, live and public.
Real products, not mock-ups: open them and try the numbers.
OWN PRODUCT · LIVE ↗
15 countries
TotalTax: sourced tax calculator
All 50 + DC US states and all 26 Swiss cantons. The build fails if any figure lacks a dated source.
- Calculator
- Rules engine
OWN PRODUCT · LIVE ↗
155 sofa shapes in 8 families
SitShape: configurator + marketplace
A shape configurator that fits a sofa to the buyer’s room, with 559 automated tests behind it.
- Configurator
- Portal
9✓ Vowko: MCP tools for members' AI agents observed 2026-10-05 · codebase
8✓ Filyhost: edge rate limiters observed 2026-10-05 · wrangler config
0.1%✓ Oil Fund Ledger: max gap to the fund's reported value over 28 years observed 2026-10-05 · reconciliation gate
Also live: Vowko, where members’ AI agents act through a built-in MCP server but can never vouch for anyone; Filyhost, a client file portal with expiring links; and Oil Fund Ledger, a public data tool whose build refuses to ship if its numbers drift from the fund’s own reports, the same gate a pricing or ROI tool needs.
10 Price
What moves the price, and what it costs to run.
Starts at €8,900 · $9,800; most tools land at €9k–25k. A price range in writing within 48 hours of our call. The fixed price follows the one-week spec, which is part of the project.
Costs more when
- Many rules, jurisdictions or product variants.
- Accounts, roles and a client portal.
- Payments and order handling.
- MCP or WebMCP access for AI agents.
- Integrations with CRM, ERP or quoting systems.
Costs less when
- Logic already working in a spreadsheet you trust.
- A public tool with no accounts.
- One language and one currency.
- One person on your side who owns the rules.
Monthly running costs
Paid by you, directly to each provider, in your own accounts. I add no markup and no hosting fee.
| What | How it works | Paid to |
|---|---|---|
| Hosting | Cloudflare. Small volumes fit Cloudflare’s free plan; Workers Paid is budgeted when volume or CPU needs it. | Cloudflare |
| Payments | Stripe’s standard fees, only if the tool sells. | Stripe |
| AI model usage | Only if the tool calls a model (to explain a result or draft a quote), capped by a monthly budget. | Anthropic / OpenAI |
11 Handover
What you own when we're done.
- Source code and test suite your GitHub organisation
- App, database and secrets your Cloudflare account
- The spec, with every rule and its source docs/ in your repo
- Domain, analytics and admin access your accounts
- Runbook for updating rules and sources docs/
No licence fees to me, no lock-in, no proprietary platform. If I disappeared tomorrow, any competent developer could pick it up from the runbook.
12 Questions
Straight answers to fair questions.
Does a calculator actually need AI?
Usually not in the calculation itself. The logic should be deterministic and tested. AI is useful around it: explaining a result in plain language, drafting the follow-up quote, or letting an AI agent query the tool on a buyer’s behalf.
What is an MCP server, and do we need one?
The Model Context Protocol is a standard way for AI assistants and agents to use a service: look something up, get a quote, start a booking. If your buyers are starting to delegate research to agents, an MCP endpoint lets those agents use your real numbers. If they aren’t yet, skip it; it can be added later.
How do you make sure the numbers are right?
Every rule is written down with its source before it’s coded, then tested against your own examples. On TotalTax, ~1,100 automated tests run on every change, and the build refuses to ship if a figure lacks a dated source.
Could AI agents misuse it?
Agents get scoped access only: limited actions, rate limits, a spending cap and a log of everything they do. Anything that commits you or your client to something can be marked “human only”, the way Vowko lets agents act but never vouch.
Who maintains it when the rules change?
Rules live in data, not code, so updating a rate or a price is an edit, not a rebuild. Your team can do it with the runbook, or I can through the AI Department retainer.
Rather talk it through first?
20 minutes with me, not a sales rep. You leave with a clear next step — whether or not we work together.