Skip to content
English
  • There are no suggestions because the search field is empty.

What are the hardware requirements for installing DeepVA on-premises?

DeepVA runs on a modular system where performance can be scaled by adding more worker nodes—ideal for real-time or parallel processing—while a minimal setup is sufficient for slower, sequential analysis with lower resource usage. 

DeepVA can be installed on-premises to meet specific privacy, security, or infrastructure requirements. The platform is built on a modular and scalable architecture, allowing performance to scale horizontally by adding dedicated worker nodes depending on workload, concurrency, and real-time requirements.

The system is composed of a Base System that runs all core platform services, and one or more Worker Nodes that execute compute-intensive AI workloads. Different worker types are used depending on the AI modules being deployed.

Contact us:
If you have questions or need guidance selecting the best setup for your workflow, our team is happy to assist—just reach out to us anytime!


Base System Requirements

The base system runs the core platform components, including:

  • Webserver

  • Backend services

  • Databases

  • Message broker

Minimum Requirements

  • CPU: 8 Cores

  • Memory (RAM): 16 GB

  • Storage: 512 GB Disk

Recommended Requirements

  • CPU: 16 Cores

  • Memory (RAM): 32 GB

  • Storage: 512 GB Disk

 

These resources ensure stable operation of the platform and reliable orchestration of worker nodes, especially in multi-worker or production environments.


Mining Worker Requirements

Worker nodes perform the compute-heavy analysis tasks. The number and type of workers can be scaled based on throughput, latency, and module requirements.

Mining Worker

  • CPU: 8 Cores

  • Memory (RAM): 16 GB

  • Storage: 512 GB Disk

  • GPU: NVIDIA GPU with minimum 12 GB VRAM

Used for standard AI analysis, extraction, and mining workloads.


LLM Worker Requirements (needed for Visual Understanding)

LLM workers are required for advanced visual understanding and large-model inference.

Mining Worker Recuirements

  • CPU: 8 Cores

  • Memory (RAM): 16 GB

  • Storage: 512 GB Disk

  • GPU: NVIDIA GPU with minimum 24 GB VRAM

Required for high-capacity models and vision-language workloads.


Scaling Guidance

  • High-Performance / Real-Time Use Cases
    → Deploy multiple worker nodes with GPU acceleration

  • Lower Throughput / Sequential Processing
    → Use fewer workers, scaling up as demand increases

The modular design allows you to mix worker types and scale independently as your workloads evolve.

 

Contact us:
If you have questions or need guidance selecting the best setup for your workflow, our team is happy to assist—just reach out to us anytime!