Skip to main content

Editorial methodology

This page explains how RivoCalc's content is produced, reviewed and corrected. We believe transparency about our process is the foundation of trust with you.

Who we are

I'm Luís Amaral, a software engineer, and RivoCalc was born in May 2026.

The short version: I needed to work out how much paint a wall would take, and went looking for something built for the UK. What I found were machine-translated foreign sites, formulas built around the wrong standards, and examples full of products you can't even buy here. Something obvious was missing: a calculator aggregator built properly for the UK, with local context — our own units, our own standards, our own way of talking about things.

So I decided to build it. Solo, with the help of AI where it scales, and with the same engineering rigour I'd apply to any commercial product: deterministic tests for every formula, multi-model content review, official sources cited on every page, zero sign-ups and zero dark patterns.

RivoCalc is independent. No investors, no sponsors, no brand affiliations. We don't sell products. We don't collect personal data. There's no sign-up needed to use a calculator.

How we create our content

RivoCalc's content is produced with the help of artificial intelligence and reviewed by a multi-model system. This approach lets us cover a large number of calculations across several markets with a consistent standard of quality.

Production process

  1. Technical specification. We define the mathematical formula, the input parameters, and usage examples. Formulas are based on official standards, technical manuals, or basic engineering calculations.
  2. AI-assisted generation. We use language models (Anthropic's Claude family) to generate the explanatory text, examples and FAQs. The AI is given the technical specification along with explicit rules on vocabulary, units and local context.
  3. Multi-perspective review. The generated content is assessed from four independent angles:
    • Technical accuracy — checks that the formula has been applied correctly.
    • SEO optimisation — checks meta tags, heading structure and schema.org markup.
    • Local relevance — checks that the vocabulary, units, currency and regulations match the target market.
    • Editorial consistency — checks that the text follows the internal standards described on this page.
  4. Independent model cross-check. After the four-perspective review, a different language model runs a cross-check and flags any inconsistencies.
  5. Deterministic validation of the calculations. For every calculator, we run test cases with known values. If the calculator doesn't return the expected results, publication is blocked.
  6. Publication. The calculator is only published once every one of these checks has passed.

Why this approach?

We rely on AI for a practical reason: it lets us produce content at a pace a solo team could never match on its own. We recognise that AI can get things wrong — it can invent regulations, mix up units, or produce incorrect calculations. That's exactly why the multi-model review system exists.

Official sources consulted

Every calculator explicitly cites the official sources behind its calculations. For the UK, our main sources are:

Sources for our other markets are still being validated. This page is updated as each market goes live.

Verification and corrections

Despite the multi-model process, mistakes can still happen. If you spot a calculation error, awkward wording, or outdated regulations, let us know through the contact form.

The form only asks for your name and email — we don't create accounts, we don't send newsletters, and your email is used solely to reply about what you've reported.

We commit to acknowledging receipt, investigating and fixing calculation errors within 30 working days, and updating figures that depend on annual tables (tax bands, regulations) as soon as they're published.

Transparency about AI

RivoCalc uses generative AI as part of its production process. We believe this is compatible with useful, trustworthy content provided the process is transparent and the output is checked.

  • The explanatory text, examples and FAQs are generated with the help of AI.
  • The mathematical formulas are defined by a human editor, not by the AI.
  • The final review runs across multiple models, to catch errors that a single model might miss.
  • Every calculator explicitly identifies the official sources it's based on.

We don't hide our use of AI. Misleading readers about where the content comes from would go against everything this project stands for.

Editorial independence

RivoCalc has no sponsors, doesn't take money from brands, and doesn't promote products in exchange for payment. Where our calculators refer to retailers or suppliers, we use generic descriptions rather than naming specific brands.

Any future monetisation of the project will be through contextual advertising only (display ads via an ad network), clearly labelled as advertising and kept separate from editorial content.

Last updated: 23 May 2026.