Core Platform Capabilities​​

AI-Powered Visual Builder

Design enterprise UIs with an AI-assisted drag-and-drop canvas. Describe what you need in plain language or design visually – PIES AI generates screens, layouts, and widget configurations that follow your organisation’s standards. 20+ professional widget types, real-time WYSIWYG editing, multi-screen support, and responsive design.​

​

AI-Powered Code Generation

Describe your application in plain English. PIES AI understands your intent and our organisation's standards, then generates complete, production-quality code.​​

Generation Capabilities

  • Complete App Generation - full apps from a single prompt
  • Smart Screen Design - AI-designed layouts with proper widgets​
  • Automatic Business Logic - event handlers wired together​
  • Database Schema Generation - data models from natural language​
  • Context-Aware Modifications - full app context for changes​
  • Multi-Stack Output: React, Angular, Go, Python, Java​
  • Real Docker Container Previews - not simulations​
  • Multi-Database: MySQL, PostgreSQL, Zetaris virtualisation​

Application Dashboard

Unlimited applications across workspaces and teams

Multi-Workspace​ Organise by team, project, or client
Applications​ Built and managed from one platform​​
Self-Hosted​ Runs entirely on your infrastructure
Private AI​ Zero data leaves your network​

DEPLOY WHEREVER YOUR ENTERPRISE INFRASTRUCTURE LIVES

Infrastructure Flexibility

Choose the model that matches your infrastructure and performance requirements. Scale up as your needs grow – from a development server running Humble Pie to a production cluster running Slab. All models share the same API, the same training pipeline, and the same data privacy guarantees. But what makes the PIES LLM family truly powerful is that you train them on your own data.

  • NO ONGOING API FEES
  • NO PER-TOKEN BILLING
  • NO DATA EGRESS
  • YOUR INFRASTRUCTURE
  • UNLIMITED INFERENCE.
On-Premise GPU Server

On-Premise GPU Server (NVIDIA A10G+)

Vertical
Private Cloud

Private Cloud (AWS, Azure, GCP)

On-demand
Kubernetes GPU Cluster

Kubernetes GPU Cluster

Horizontal
Air-Gapped / Secure Enclave

Air-Gapped / Secure Enclave

Fixed
Development / Edge Server

Development / Edge Server

Minimal

PIES AI Administration Portal

Manage, train, and control your private AI from a dedicated dashboard

It learns your organisation’s vocabulary, your preferred UI patterns, your data model conventions, your business logic structures. Over time, it generates applications that feel like they were built by someone who has worked at your company for years.

The more your team uses the platform, the better your private AI gets – automatically.

Training Data Management

Training Data Management

Upload, review, and curate JSONL training datasets

Training Control

Training Control

Configure epochs, batch size, learning rate, and LoRA rank

Live Training Monitor

Live Training Monitor

Track training progress in real time (0-100%)

Adapter Management

Adapter Management

List, activate, or roll back trained model versions

Playground

Playground

Test your trained model interactively before deploying it

Deploy wherever your enterprise infrastructure lives

Complete Data Sovereignty

Your prompts, your training data, your generated code, and your model weights never leave your infrastructure.

Data sovereignty applies when using PIES Private LLMs. Optional cloud providers (OpenAI, Anthropic) are subject to their respective policies.

Data residency

Data residency

All inference runs locally. Zero external API calls.

Training data ownership

Training data ownership

Stored on your file system. Never transmitted. Never shared.

Model weights

Model weights

Run on your hardware. No cloud dependency.

Air-gapped deployment

Air-gapped deployment

Fully functional without any internet connectivity.

Audit trail

Audit trail

Every training run timestamped. Adapters versioned and stored.

Data deletion

Data deletion

Delete training data and adapters at any time via admin portal.

Access control

Access control

PIES AI admin restricted to superusers only.

Enterprise-ready from day one

Security & Compliance​ ​

Additional AI Providers (Optional)​​

For organisations that want the broadest capability, PIES Studio also integrates with OpenAI GPT-4o and Anthropic Claude as supplementary options alongside your private PIES LLMs. Your intelligent router directs requests to the right model. Cloud providers are entirely optional - PIES Studio is fully functional with PIES LLMs alone.​​

Authentication

JWT + RSA-2048 key management​

Authentication

Password Security​

Bcrypt hashing​

Password Security​

License Encryption​

AES-256-GCM​

License Encryption​

Secrets Management​

HashiCorp Vault integration

Secrets Management​

Access Control​

Role-based (Admin, Developer, Viewer)​

Access Control​

Generated Code Security​

SQL injection prevention, CORS, CSRF protection​

Generated Code Security​

AI Data Sovereignty​

Local inference, local training, no external calls​

AI Data Sovereignty​

Offline Capability​

Fully air-gapped deployment support​

Offline Capability​

Multi-Stack Code Generation

Enterprise-ready from day one

Live Preview System

See your application running in real time. PIES Studio spins up real Docker containers – not simulations – giving you an authentic preview as it will behave in production. Instant preview after generation, real-time build progress, hot reload, and automatic container lifecycle management.

Multi-Database Support

Connect to your data wherever it lives. Native integration with MySQL and PostgreSQL. Zetaris data fabric for unified access across any JDBC source. Oracle and SQL Server planned. Real-time schema introspection, direct data model visualisation, and cross-database relationship support.

Frontend

React / Next.js React / Next.js
Angular 19 Angular 19

Backend

Go (Chi) Go (Chi)
Python / FastAPI Python / FastAPI
Java / Spring Boot Java / Spring Boot

Every generated application includes full-stack scaffolding, database integration, typed API routes, Docker containerisation, and production-ready code you fully own and control. Extract, modify, and deploy independently – no runtime dependency on PIES Studio.

Architecture

Modern microservices - reliable, scalable, and flexible​

  • PIES Studio IDE
  • Design
  • Generate
  • Debug
  • Deploy
  • Manage

Enterprise Infrastructure:

Database . Cache . Secrets . Containers

  • Studio IDE: Visual design canvas, AI chat, debugger, deployment management​
  • Core Orchestrator: Central API gateway, authentication, real-time communication​
  • AI Generation Engine​: Central API gateway, authentication, real-time communication​
  • PIES Private LLM​: Organisation-trained model, local inference, zero data egress​
  • Deployment Engine: App building, containerisation, cloud deployment, SSL/TLS​
  • Preview Runtime: Live application preview with real containers and hot reload​​
  • Database Connector​: Multi-database access and data virtualisation
Core API &
Orchestrator
PIES AI
Generation
Engine
Deployment
Engine
PIES
Private LLM
Engine
Container
Orch.
& Preview

PIES Studio Platform​​

System Requirements

Minimum

(Development / Evaluation)

  • CPU

    4 cores

  • RAM

    16 GB

  • STORAGE

    50 GB free

  • OS

    Linux (Ubuntu 20.04+, Debian 11+, RHEL 8+) or any Docker-supported OS

  • CONTAINER RUNTIME

    Docker Engine 24.0+ with Compose v2

Recommended

(Production)

  • CPU

    8+ cores

  • RAM

    32 GB+

  • STORAGE

    100 GB+ SSD

  • OS

    Linux (enterprise distribution)

  • CONTAINER RUNTIME

    Docker Engine 24.0+ or Kubernetes 1.26+

  • NETWORK

    Internal bridge network between services

Start Building Your AI-Powered Enterprise Application

Transform Your Requirements into Working Applications

Book a Demo