Choose sources carefully
Research, public guidance and marketing can address the same subject without carrying the same weight.
Learn more
Digital evidence methods, applied to nutrition
Nutrition needs evidence, not noise.
We envision specialist AI systems that organise human and animal nutrition information, trace where claims come from and explain what remains uncertain. Technology should make understanding easier, not replace judgement.
Identifiable documentsContext-aware answers
Quality-led engineering
Good answers begin before the AI.
Before generating a sentence, a system needs rules for the documents it searches, the claims it cannot substantiate and the uncertainty it must report. These decisions shape the product.
Show what is actually known
Interfaces can expose references, clarify their scope and admit that an answer is incomplete.
Learn moreKeep people in control
Users gain the tools to navigate material and ask more precise questions.
Learn moreTwo domains, shared standards
From human nutrition to petfood information
Different kinds of content call for specialist tools, with the same commitment to understandable provenance.
Examine nutrition claims
Browse general topics, find source material and learn why some answers require qualifications.
Human nutritionLook beyond the ingredient list
Make petfood labels, product descriptions and supporting references easier to inspect.
PetfoodInterface preview
Evidence should keep its context.
This fictional interface demonstrates how reference categories and limitations might appear in a future documentary tool. It is not a live scientific analysis.
DocumentationQuestion: what supports this nutrition claim?
The workflow locates relevant material, classifies document types and highlights questions still requiring verification.
Source document identifiedFor review
Scope needs checkingCheck
Conclusion needs contextCheck
Question: how should a petfood label be read?
The workflow separates ingredients, declared values and promotional content without judging suitability for an individual animal.
Product labelFor review
Comparable measuresCheck
No individual adviceCheck
From requirement to product
A four-stage development process
Frame
Define the real use case and intended audience.
Ground
Identify permissible sources and terms of use.
Build
Implement interfaces, automations and checks.
Test
Evaluate output quality and document limitations.
Information is not medical treatment. These are software and educational concepts. They do not replace medical nutrition consultations, dietetic care or veterinary advice.
Experience rooted in digital investigation
Cyberdétective: from investigative methods to software tools.
After an early career specialising in computer investigations, Alain Stevens moved into software engineering, IT advice, SEO, AI applications and training. OSINT tools are designed for users to operate themselves; no named private investigation service is undertaken for third parties.
Read the backgroundPacta Labs
Pacta Labs develops software concepts, documentary systems and cross-industry digital services.
Explore ↗OSINT tools
Cyberdétective software for OSINT workflows, fraud prevention, cyber awareness and self-directed research.
Explore ↗Press & media
Professional background, Cyberdétective history and media enquiries about AI, information quality and cybersecurity software.
Explore ↗Pricing
Pricing approach for IT consulting, nutrition AI prototypes, web applications, software and technical training.
Explore ↗