Two ways to run Deep Netts
Use the library on your own servers, or let us run it for you on AWS and pay by the hour.
Self-managed
Deep Netts is a Java library. Add it to your project with Maven and it runs wherever your application runs: a laptop, your own servers or your own cloud account. Your data stays with you.
Who does what
- Servers, cloud account or cluster
- Your Java applications and their deployment
- Model training data and model files
- Security, networking and compliance
- The Deep Netts library and its updates
- Your licence file
- Documentation, examples and course
- Support and SLA (Enterprise)
Where it runs
- Windows, Linux and macOS with JDK 25
- Docker and Kubernetes
- Your AWS, Azure, Google Cloud or Oracle account
- On-premises data centers and private clouds
- Servers with NVIDIA GPUs
- Desktop applications
What you get
A Maven dependency – no separate AI service to deploy or monitor.
A signed file checked locally – no licence server, no call to us.
The JDK Vector API uses every vector lane of your CPU (Free and Enterprise).
Training and inference with CUDA, cuBLAS and cuDNN (Free and Enterprise).
Load the model once, call predict() from a service.
Design and train networks visually, then use them from Java (Free and Enterprise).
When self-managed is the right choice
Fraud detection, payment routing and risk scoring, on-premises.
Customer, patient or research data that must not leave your environment.
No internet access needed – the licence is checked from a file.
Ship models inside your own Java software.
Managed by Deep Netts
We run Deep Netts for you on AWS. You train and run your models; we set up the servers, keep them updated and bill you for the hours you use.
Who does what
- Your data and use case
- Your AI applications
- Choosing the instance size
- Installation and configuration
- Updates and patches
- Monitoring and scaling
- Licence and basic support
Where it runs
- Our AWS cloud
- Inside your own cloud (VPC) – on request
What you get
Everything in Enterprise Edition, run by us.
Vector API and CUDA, sized to your workload.
Only the hours training and inference run, on the instance size you choose.
Included with the service.
When managed is the right choice
Start without buying GPU servers or building a platform team.
Pay for the hours you use, scale down afterwards.
Try a use case end to end before deciding how to run it.
On request we run the platform inside your VPC.
Side by side
Where Deep Netts runs, who operates it and what each edition includes.
Same library in both. Same Java API, same model files – switch either way.
Self-managed Community | Self-managed Free | Self-managed Enterprise | Managed by Deep Netts Managed Service | |
|---|---|---|---|---|
| Where it runs | Your infrastructure | Your infrastructure | Your infrastructure | Our AWS cloud (your VPC on request) |
| Who operates it | You | You | You, with our support | Deep Netts |
| Where your data is | Your environment | Your environment | Your environment | Our cloud or your VPC |
| Acceleration | Basic CPU | Vector API + NVIDIA GPU | Vector API + NVIDIA GPU | Vector API + NVIDIA GPU |
| Production use | Under the GPL – your code stays open | One environment, product under $100k a year | Unlimited | Unlimited |
| Support | Community (GitHub) | Community | Enterprise support + SLA | Basic included, SLA |
| Price | Open source | $0 | Custom | Per compute hour |
Which edition do I need?
The heart of Deep Netts, developed in the open. Clone it, read it, build with it.
- Pure Java deep learning engine
- Open core of the Deep Netts platform
- Free for learning, research, and building
- Community support on GitHub
The full accelerated platform, free to run in one real production deployment while your product is small.
- Everything in Community Edition
- CPU acceleration via the Vector API
- GPU acceleration via CUDA
- Visual AI Builder
- Source code and license included in the download
- – One production environment
- – Annual revenue from the product using Deep Netts under $100,000
- – Not offered as a hosted AI platform, managed service, or SaaS
Deep Netts run for you on our infrastructure – training and inference billed like cloud compute, by the hour. No license, no setup, no ops.
- Everything in Enterprise Edition
- Runs on our infrastructure – nothing to install
- Pay per compute hour, like a cloud instance
- Scale training and inference up or down
- Updates, monitoring, and operations handled by us
Installed in your own environment under an enterprise license. Scale and support for companies whose product earns more than $100,000 a year.
- Everything in Free, without revenue limits
- Multiple production environments
- SaaS and hosted deployment rights
- Technical support and maintenance
- Service-level agreements (SLA)
Questions about running and licensing Deep Netts
What's the difference between Community and Free?
Community Edition is the open source core – the pure Java engine, developed in the open on GitHub. The Free edition adds the accelerated engine (Vector API on CPUs, CUDA on GPUs) and the Visual AI Builder, free for one small production deployment.
Can I evaluate Deep Netts for free?
Yes. Development, testing, and evaluation are always free – the conditions of the Free edition only apply once you deploy to a production environment.
Managed Service or Enterprise Edition – which one is for us?
Both give you the full platform without limits. Enterprise Edition is a license: you install and run Deep Netts on your own servers or cloud account, and we support it. Managed Service is the same platform run by us on our infrastructure – you pay per compute hour, like a cloud instance, and never touch the setup.
How is the Managed Service billed?
Like cloud compute. You pay for the hours your training and inference workloads actually run, on the instance size you choose. There is no license fee and no minimum commitment – write to us and we will size it with you.
Can the managed platform run inside our own cloud?
On request. For teams that need the data to stay in their own VPC, we can run and operate the managed platform inside your private cloud – your infrastructure, our platform and operations. Contact us for details.
Where can I read the full license terms?
The Community Edition license is in the GitHub repository. For Free and Enterprise, the End User License Agreement ships with every download and is the authoritative document. If anything is unclear, write to contact@deepnetts.com.