What is Perfact, and how does it work?
Perfact helps people make sense of food and product labels without having to decode every ingredient and number themselves. We take ingredient and nutritional data, along with additional information not found on the label – such as consumer ratings and regional regulatory approvals – organise it, and let users assess products against the criteria that matter to them, including specific diets, ultra-processing, vegan standards, allergens and animal testing.
In short, we’re turning complicated, impractical product data into clear choices people can make in the aisle, at home, or online. For example, there are more than 250 different names for sugar on food labels, which is overwhelming for someone in the supermarket looking to avoid added sugar. Perfact will give you a thumbs-up, a thumbs-down or recommend alternatives, within a fraction of a second.
How did Perfact come about?
I’m a computer scientist driven by a passion to use technology to make life simpler and more manageable for people. After a girlfriend introduced me to Wagner, I became fascinated with the patterns, motifs and structures in classical music that experts could hear and I couldn’t, and wanted to build a tool to unlock the complexity and make it visible and accessible for all.
That took me to Stanford and a PhD, building a prototype for that tool to apply pattern analysis across a vast music database. Then, when speaking with a friend who lived in shared accommodation with someone who had allergies, I realised that food labels gave people lots of information but very little practical help to enable them to buy and eat with confidence. That consumer-facing database challenge then took me to the Stanford Graduate School of Business, and sowed the seeds for Perfact.
How do you gather the data?
The data comes from several different sources. We use public databases, such as USDA FoodData Central, although the raw data often needs a considerable amount of cleaning up. In Europe and the UK, we typically access product information through retailers, who publish the ingredient and nutrition data supplied by manufacturers for online shopping. The collection process is automated, though we don’t simply dump the raw data into the system – the datasets we’re dealing with are so vast, scientific, complex and variable that AI is unable to process them with acceptable error rates at this point. We validate, clean and structure the data, and then consistently apply our filters and matching logic on top to ensure users can trust the outcome.
We’re simply making food labels transparent and diets easy for people to follow
The question of whether a particular food or diet is healthy or unhealthy is highly contentious. How do you decide?
We don’t – we’re the messenger in this dynamic. We’re simply making food labels transparent and diets easy for people to follow. The principles of the high-fat, low-carbohydrate Keto diet are very different from a scientific framework such as Nova, which focuses on how a food has been processed, which in turn is very different from the Metabolic Matrix, which is concerned with how products impact the metabolism. Our role is not to say whether a diet is ‘healthy’ – it’s to make the criteria actionable and transparent for users who want to follow it.
Whilst other sites and apps out there may apply valuations to products based on their own, often opaque, assumptions and criteria, we restrict ourselves to letting people individually combine any number of fully transparent criteria. Similarly, we don’t allow ‘offsets’ – we just show users whether a product complies with a diet of their own choosing.
We do, however, label filters that are grounded in scientific evidence. Other filters can exist on the platform too, including ones based on personal beliefs or lifestyle views, but they don’t get that academic marker. That way, people can tell the difference, and we stay neutral.
How can people find Perfact?
The Perfact.co website, which works in a mobile or desktop browser, is live, and the US version of our app will be live in the next couple of months, meaning anyone will be able to scan a product barcode when shopping and know immediately whether it complies with a particular diet and how it compares nutritionally to the alternatives.
Ultimately, however, our ambitions for the scope of the platform go well beyond consumer-facing applications. Supermarkets or delivery platforms that currently rely on manual tagging could use our API to improve online filtering and provide recommendations or alternative products with confidence, whereas manufacturers can use the tool to test or reformulate products and see whether they meet the requirements of a particular diet or health framework. And to our delight, that’s already happening.
In addition, groups such as WWF or PETA could use the platform to turn their own criteria into something practical for consumers, without building and maintaining an app themselves. We’re already doing this for the Hypoglycemia Support Foundation.
How much does it cost?
Anyone can use Perfact anonymously and at no cost to search products, apply filters, build a personalised profile, and organise items into a list that can be transferred to Amazon or Walmart shopping carts. Registering with an email address ensures those profiles and lists are saved permanently and can be accessed from any device.
For $10 a year, subscriber benefits include enhanced tools for organising and saving products, input on which filters we build next, and the chance to attend live events with experts. An additional $5 per year unlocks features such as extra capacity to save and manage more products, priority entry to live sessions, and access to recordings of past events and premium content from list creators.
Dr Andreas Kornstädt was interviewed for Metabolica on 2nd April 2026
