In today’s media landscape, artificial intelligence is often presented as something complex, abstract, and difficult to integrate. But what if AI became nothing more than a function – embedded directly into your workflows, invisible yet powerful?
That’s exactly what the partnership between DeepVA and helmut.cloud delivers. DeepVA brings cutting-edge composite AI for audiovisual understanding, while helmut.cloud provides the collaboration platform that makes AI accessible exactly where it’s needed. Together, they allow broadcasters, media houses, and corporates to turn AI into their AI.
High5: Bringing AI into your favorite Stream Designer
The integration of DeepVA into helmut.cloud is realized via High5, the workflow orchestration engine at the heart of helmut.cloud. With High5, workflows – or “streams” – can be created visually in the Stream designer Studio. From no-code to low-code or advanced scripting, High5 lets users design, monitor, and optimize their workflows seamlessly.
- Open your workflow
- Add in the DeepVA nodes
- Use AI in your workflows
As simple as that – and you are ready to understand not only generic context, but also regional knowledge based on your own custom ai models and datasets, deeply tied into your daily workflows. This ensures flexibility, security, and control, while keeping workflows frictionless for the user.
Examples of a Smart Workflow in combination of DeepVA and High5
Use Case: Automated Shotlists
With an influx of thousands of minutes of new content per day, big media houses need to keep an overview of what is relevant. This use case automates the ingesting of new assets and delivers not only metadata, but a rich, contextual overview as shotlist, even for regional media outlets.
- New Videos are added to a watch folder
- The assets are automatically sent from the ingesting folder to DeepVA.
- Based on predefined nodes, multiple analyses are triggered: recognition of regional faces and voices, identification of local landmarks, and extraction of general metadata such as transcription, translation, and scene-based content summaries.
- All separated metadata is sent back and becomes instantly available, also for further workflows
- The retrieved metadata is parallel processed and combined into a timecoded shot list that is even applicable to regional content, saving valuable time and overcoming the limitations of generic AI solutions.
- Our collaboration remains behind the scenes. And the results are delivered straight where your users need them.
Result: Richer metadata, easier & faster access, insights for automated editing and a more productive journalism at scale.
Sovereign & Composite AI – why it matters
Generic hyperscaler AI services often create more problems than they solve. Large Language Models struggle with regional people, voices, or landmarks. Their token-based billing makes costs unpredictable. And worst of all, customer data is frequently reused for training.
DeepVA and helmut.cloud take a different path: Sovereign & Composite AI. With one API, you can mix open source, proprietary, and third-party algorithms into workflows that fit your needs. Everything runs either on-premises or in the cloud, fully GDPR-compliant and independent of geopolitical shifts. Most importantly: your data stays yours – DeepVA never uses it for training.
This approach also enables true localisation. Instead of relying on generic models, you can train AI to recognise regional faces, voices, and landmarks, and enrich your content with metadata that resonates with your specific audience.
From collaboration to context: AI that works for you
Unlike hyperscalers, DeepVA and helmut.cloud offer a combination of:
With DeepVA x helmut.cloud, AI becomes invisible yet indispensable:
Fast, secure, flexible and tailored to your workflows.
Turn AI into your AI.
