Skip to content
    Mandatory publicity

    The MONOKAIDO project and the Recovery Plan

    We completed the last research and development phase of the Monokaido platform with support from Slovakia's Recovery and Resilience Plan. On this page you will find what the subject of the investment was, what goals the project set itself and how its results were measured.

    Co-funded by the European Union — NextGenerationEU with funds from Slovakia's Recovery and Resilience Plan.

    EU emblem with the Slovak wording for “Funded by the European Union — NextGenerationEU”Logo of the Recovery and Resilience Plan of the Slovak Republic with the Slovak wording “Plán obnovy” (“Recovery Plan”)
    Basic information

    Project identification

    Recipient
    For Best Clients, s.r.o., Trnava
    Project title
    MONOKAIDO – an AI-powered SaaS platform for product information management (PIM) and e-commerce with integrated generation of content, translations and visuals
    Internal project code
    monokaido-q2
    Implementation period
    May–June 2026, completion deadline 30 June 2026
    Thematic areas
    Key digital technologies · Artificial intelligence and robotics
    Source of funding
    Slovakia's Recovery and Resilience Plan (Plán obnovy a odolnosti SR) — NextGenerationEU
    About the project

    From a prototype to a system ready for deployment

    The aim of the project was to move the platform from technology readiness level TRL 6 to TRL 8 — a complete, qualified system ready for pre-commercial demonstration with pilot customers.

    A mid-sized online shop today manages tens to hundreds of thousands of items, imports data from dozens of suppliers in mutually incompatible formats and runs several language versions in parallel. Product managers spend weeks writing descriptions, SEO content is bought in externally and connecting every new supplier feed takes a development team four to six weeks. This operating model pushes the cost of handling a single catalogue item to a level at which smaller brands lose their competitiveness.

    Monokaido is a multi-tenant SaaS platform for product information management and e-commerce that combines a proven technology stack (NestJS, Go, PostgreSQL with Row-Level Security, Redis, Next.js) with a natively integrated layer of generative and multimodal artificial intelligence. At the core of the innovation are five closely linked AI modules that address the five most expensive manual processes in the real-world operation of an online shop: orchestration of multiple AI providers, extraction and normalisation of parameters, generation of SEO content, product photography and video, and the connection of supplier feeds.

    The project falls within the thematic areas “key digital technologies” and “artificial intelligence and robotics” — both are explicitly supported by Slovakia's Recovery and Resilience Plan.

    Project goals

    Five goals and their measurable results

    Every goal has a result that can be verified with a number — not a declaration.

    1. 01

      Multi-provider AI architecture with automatic fallback

      • Integration of five AI providers with automatic switching when a rate limit is hit, and a target AI service availability of 99.5% at a consumption of at least 1 million tokens a month.
      • Launch of the platform-managed provider Monokaido AI with a tier-based token bonus as part of the subscription.
    2. 02

      Validation of AI parameter extraction and normalisation

      • Parameter extraction from product photos and texts with an accuracy of ≥ 92% on a validation set of at least 10,000 products.
      • AI mapping of parameter names and values with deduplication across supplier feeds.
    3. 03

      Content Generator with dynamic internal linking

      • Generation of SEO content for categories, products and articles with {linkN} placeholders that are resolved per language only at publication.
      • A job-based Generate → Review → Apply architecture with a full audit trail and recovery of stuck jobs.
    4. 04

      AI Photo Studio and generation of product videos

      • Background removal, scene generation and product videos with an average response time of up to 30 s for an image and up to 300 s for a video.
      • AI generation of ALT texts and SEO attributes for at least 95% of the images in the catalogue.
    5. 05

      AI dynamic feed mapping for supplier onboarding

      • Onboarding of a new supplier into production within 15 minutes of administrator work instead of the original approximately 6 weeks of development, with at least 10 supported formats.
      • A state model of the mapping source, manual ↔ ai_generated → ai_edited, with auditable changes and SSRF protection on every URL input.
    Implementation process

    Six work packages in two months

    The development packages ran in parallel in May 2026, the pilot operation and final preparation in June 2026.

    PB01May–June 2026

    Project management

    Management, coordination, monitoring and control of the project's implementation, including the process set-up.

    PB02May 2026

    Development of the AI core and the multi-provider router

    An AI orchestration layer with cooldown and fallback, Monokaido AI with a tier-based bonus, an audit log of consumption and an admin dashboard.

    PB03May 2026

    AI parameter extraction and normalisation

    A pipeline for extracting parameters from photos and text, mapping of names and values across suppliers, and validation on a production dataset.

    PB04May 2026

    Content Generator and AI dynamic feed mapping

    An automatic Content Generator with internal linking, and the AutoAdd vendor flow with an AI generator of feed mappings.

    PB05May 2026

    Photo Studio, video and multilingual translations

    A unified Photo Studio interface with provider routing, and completion of the bulk translation pipeline.

    PB06June 2026

    Pilot operation, business strategy and documentation

    Deployment with pilot customers, evaluation of the measurable results, documentation and go-to-market preparation.

    Expected benefits

    What the project brings to customers and the market

    Lower product content costs

    Automating descriptions, translations, photos and ALT texts reduces the operating costs of content by roughly 60 to 80% compared with a manual approach.

    Faster supplier onboarding

    AI analysis of the supplier feed shortens the connection of a new supplier from six weeks of developer work to minutes of administrator work.

    Higher organic traffic and conversion

    Generated SEO content with consistent internal linking and better product pages increase both the traffic and the conversion rate of an online shop.

    Stronger cyber security

    A multi-tenant architecture with Row-Level Security, encryption of API keys, SSRF protection on supplier feed URLs and an audit log of AI calls for every company.

    Resilience to the outage of a single provider

    A router with automatic fallback between five AI providers removes the dependence on a single model supplier.

    A stronger position for Slovak SaaS in Central Europe

    Completing the platform opens the way to expansion into the Czech Republic, Hungary, Poland and Austria, where there is comparable demand for localised PIM solutions.

    Let's try it on your own catalogue

    We'll walk through your feeds, categories and parameters live.