Our products
They all run today on critical data and real work, across domains that share nothing in common: insurance, finance, restaurants, healthcare and engineering. At the center, our flagship: Naktor ARGOS. Around it, a catalog already working inside companies.
In Production, Not a Demo
No demos, no proofs of concept. Every product on this page runs inside a company today, with its real data and its real consequences.
Proprietary Architecture
Designed and built by us, from the infrastructure to the model. Not a wrapper over someone else's API — we build what's underneath, not just the interface.
Built for Your Security Team
On-premise data where the sector demands it, and GDPR and EU AI Act compliance where it matters. Built to pass the security review, not just the product demo.
Naktor ARGOS
The full development cycle, with agents at every phase: from requirements to QA.
The platform a team uses to design, prototype and build software with agents across the whole cycle. Born in a demanding enterprise environment, now our own product, organized into legs that cover each phase.
Requisitos
AI-assisted enterprise requirements
Turns conversations and scattered documentation into clear, traceable specifications. The leg already running in production.
Naktor Prototype
From requirement to working prototype
Takes the requirements export from Naktor ARGOS and builds a working prototype automatically, so you validate before building.
Naktor Dev & QA
Dev and QA teams, at another level
An ecosystem of AI assistants, tools, skills and techniques built to multiply what a dev and QA team can ship.
Document infrastructure
Chaotic documents turned into structured data and systems that act. The invisible plumbing a company needs before it can use AI for real.
Naktor Knowledge
Ask questions of your contracts, files and records. The originals never leave your server.
Connects your document sources, understands what's inside and lets you ask questions in natural language across the whole corpus, however large and scattered it is.
Underneath, it builds a runtime entity schema and combines semantic search with SQL over the metadata: a single question can cross thousands of documents without depending on them being well tagged. And it all runs on your infrastructure: the original files are queried in place, never copied to a third party.
- Source connectors
- Runtime entity schema
- Hybrid search + SQL over metadata
- Originals are never copied
Fororganizations with large corpora (contracts, files, invoices) that want to query them like asking a person.
Naktor Extractor
Your client's impossible PDF, turned into clean markdown, on your own GPU cluster.
Turns your most complex documents into AI-ready material. Our own pipeline on our GPU cluster, with the best OCR engines on the market.
The pipeline is our own and runs on our GPU cluster, combining CPU and GPU with the best OCR engines on the market so it doesn't choke on scans, tables or figures. It's exposed as an API the rest of the fleet reuses, so any product that needs clean documents asks here instead of reinventing extraction.
- PDF, DOCX, XLSX, images → markdown
- Hybrid CPU + GPU pipeline
- Our own GPU cluster
- API reusable across the fleet
Forany workflow that needs to feed hard documents (scans, figures, tables) into an AI system.
Naktor Invoice
The invoice arrives by email, your ERP records it. A human steps in only on conflicts.
In production and ready to connect. Plug in an agent or your inbox, it reads the invoices and integrates them automatically into any enterprise system. And when something doesn't add up, a conflict-resolution flow with a highly visual human interface.
Invoices come in through an agent or straight from your inbox, get read and posted into your ERP with no manual typing. When a field doesn't add up, instead of failing silently it opens a visual conflict-resolution flow where a person decides in seconds. It deploys as SaaS or on-premise, depending on what your sector requires.
- Read by agent or by email
- Integration with any ERP
- Visual conflict resolution (human in the loop)
- SaaS or on-prem
Forfinance and operations teams that receive many invoices and want them in their system without manual typing.
Vertical intelligence
AI built for one specific sector, not retrofitted afterwards. From a TripAdvisor listing to a medical consultation under the EU AI Act.
RestoData
LiveWhat Google and TripAdvisor know about your restaurant, turned into concrete action every month.
Turns every review into an opportunity. We analyze your reputation across Google and TripAdvisor with AI and give you actionable recommendations, monthly reports and a chat to understand what drives your rating.
Every month it analyzes with AI what's being said about your restaurant on Google and TripAdvisor and turns it into concrete recommendations, not one more dashboard to stare at. You get a monthly report on what's driving your rating and a chat to ask it directly why it goes up or down, without living glued to the platforms.
- AI review analysis
- Google + TripAdvisor
- Monthly reports
- Chat about your reputation
Forrestaurants that want to understand and improve their online reputation without living glued to the platforms.
Remor
The consultation transcribes itself and the report writes itself, under GDPR and the EU AI Act.
Transcribes the consultation in real time, generates the report and structures the clinical documentation, so the professional spends their time on the patient, not the paperwork.
It transcribes the consultation in real time, drafts the report and structures the clinical documentation while the professional is with the patient, not afterwards. It integrates with the center's HIS systems and meets GDPR and the EU AI Act out of the box, because in healthcare that isn't optional. The goal is simple: give the doctor back the time that today goes into paperwork.
- Real-time transcription
- Medical report generation
- GDPR & EU AI Act compliant
- HIS system integration
Forclinics, hospitals and healthcare professionals who want to cut administrative load.
Tools for builders
The tools we use every day to build agents, opened up to anyone building with AI too. From prompt engineering to the full map of Claude Code's internals.
SystemPromptMaster
LiveProvider-optimized prompts, exportable to n8n, Make or Zapier.
Turns a vague idea into a precise system prompt, optimized per provider and ready to export into your automation workflows. With a community template library.
You start from a vague idea and out comes a precise system prompt, optimized for the provider (it's not the same for ChatGPT, Claude or Gemini) and ready to export to n8n, Make or Zapier. On top you get a community template library so you don't start from scratch, and one place to create, improve and organize all your prompts instead of scattering them across notes.
- Provider-optimized prompts (ChatGPT, Claude, Gemini)
- Export to n8n, Make and Zapier
- Community template library
- Create, improve and organize your prompts
ForAI engineers and automation builders who need reliable, repeatable prompts in their pipelines.
Claudepedia
Live84 patterns, 63 tools, 100+ commands: everything Claude Code doesn't document, in one place.
Documents, explains and makes searchable everything inside Claude Code: its architecture, patterns, tools and safety controls. A practical map for anyone building agents.
It explains Claude Code's architecture from the inside (the agent loop, the tool system, the prompts, the safety controls) and makes it searchable by domain and tags. It also catalogs the ecosystem of tools and frameworks and adds practical guides on agents, context, permissions and debugging. It's the map we use ourselves to build agents, published for anyone else building them too.
- Architecture explained: agent loop, tool system, prompts
- Searchable pattern index by domain and tags
- Catalogued ecosystem of tools and frameworks
- Practical guides: agents, context, permissions, debugging
Fordevelopers building agent systems who need to understand Claude Code internals and apply its patterns in production.
The ecosystem, not just the tools
A product solves one concrete problem. What actually changes a company is the whole: the capabilities we install in your team, a way of working with AI, and the infrastructure that holds it up underneath. That's what we build with you — not just a standalone tool.
Ideas in our pipeline
Concepts we've prototyped or keep circling. Not in production. Yet.
Leadbreath
B2B prospecting automation: AI defines campaigns and delivers qualified meetings.
Eternity
Interactive digital memorial from photos and videos of your loved ones.
Vibeblogging
Voice to blog: you speak your ideas and AI turns them into posts in your style.
Need something similar?
These products demonstrate what we know how to do. If your company needs a similar or completely new AI solution, let's talk.