divis.io

What it is

DIVISIO is a German AI consulting and development company. It is not positioned as a general-purpose online tool, but as a partner that helps businesses apply machine learning, deep learning, and AI systems to specific business workflows. The website also showcases STAG AI Image Tagger, an open-source local image auto-tagging tool for photographers.

Core capabilities and use cases

DIVISIO covers the full lifecycle of AI projects: requirements analysis, concept design, software architecture, data collection and cleaning, data annotation, test planning, model development and training, system integration, deployment, maintenance, and employee training. Technically, it can build custom software, adapt DIVISIO products or open-source models, or integrate third-party APIs from Google, Amazon, IBM, DeepL, and others. Cases listed on the website include AI AutoIdent, vibration and thermal imaging analysis, financial text generation, product image classification, ultrasound-based railway measurement analysis, and Transformer-based automatic generation of news clippings.

STAG is a more specific tool: it uses the recognize-anything model to identify general image categories and writes them into XMP Sidecar files, making it compatible with photo management workflows in software such as darktable and Capture One. It is suitable for batch-generating tags like “horse,” “snow,” and “tree,” helping improve photo library search efficiency.

Pricing and privacy

There is no public pricing for enterprise consulting, development, hosting, or maintenance; customers need to contact sales for a custom quote. STAG is described as an open-source tool, with binaries for Windows, macOS, and Linux, plus source code on GitHub. It does not require cloud access or a subscription. On privacy, DIVISIO states that it assesses whether data can be processed by third parties, or whether it must remain in Germany or within the customer’s own IT systems. Deployment can be on customer premises, DIVISIO hardware, a cloud provider, or another service provider. STAG runs locally; apart from the initial download of an approximately 3.2GB model, image processing does not depend on the internet.

Pros, cons, and who it is for

The strengths are its complete project methodology, with an emphasis on enterprise integration, data quality, testing, and long-term maintenance; its technical approach is also relatively flexible. The downside is that the website lacks pricing, project timelines, SLA details, and performance metrics, so procurement requires thorough discussion in advance. STAG is easy to use and privacy-friendly, but it can only recognize general categories and is not suitable for faces, landmarks, specific individuals, or fine-grained species identification.

Access from China and alternatives

The website does not specify availability in mainland China, supported payment methods, or service coverage, so these should be considered unknown. Enterprise customers should confirm network access, contracts, payment arrangements, and data export requirements in advance. As alternatives, enterprise AI projects can be compared with local AI solution providers or cloud vendor AI services; for photo management, users can compare Lightroom, built-in capabilities in darktable/Capture One, and other local image recognition tools.