PIES Studio · build with AI you own
PIES Studio turns a brief into a real application: database, screens, logic, sign-in, documents, mobile app and its own AI agents. Start free on your own machine: Hobby runs its AI on your computer, nothing to pay, nothing leaves your desk. Scale up and your rules decide which model sees what, your organisation keeps what it learns, and you keep the source.
Control your IP
Most AI builders send every prompt, every record and every line they write to someone else's cloud. PIES Studio Hobby ships with PIES Private AI Micro, a model that runs on your own computer. Free, offline, and nobody else's.
Micro runs on your computer. No account to create for it, no key to paste, no per-call bill. Pull the network cable and keep building.
Your brief, your data and the apps you build never go to a cloud model unless you connect one. We test the local AI for network calls on every release: it makes none.
What you build runs on your machine with a real database, screens, sign-in, documents and a mobile app. Download the source whenever you like and run it without PIES. When you grow, Enterprise runs it on your servers.
Micro handles chat, questions over your data, in-app chatbots, small edits and building one screen, data model or event at a time. Whole-application builds use PIES Private AI Lite and Ultra or a cloud model under your policies, through Enterprise or PIES Cloud. Micro needs 16 GB of memory and an 8 GB one-time model download.
Trusted by enterprises building the future
01 · Build
Describe the app. PIES AI creates the tables, relationships, screens and logic, then shows a receipt of exactly what it changed. Undo anything with one click.
02 · Run
What you build is a working application with your data, not a prototype. Export standard source code in the languages your engineers already know, or deploy it from PIES Studio.
03 · Agents
Create an agent from a brief, give it tools and a role, post it to an environment. You decide how much autonomy it has and the app enforces it.
04 · Govern
Use the best cloud models, or PIES LLM running privately on your hardware. Policies redact sensitive data and decide which model may see what. Change a rule and every app follows it.
05 · Record
Who asked, which model answered, what it cost and what was allowed or blocked. A tamper-evident audit trail your security team can read.
06 · Operate
Promote through Preview, Test and Production to Docker, Kubernetes, AWS or Azure. Operate from one control panel: health, logs, restart, scale, roll back, schedules. Ask PIES AI to deploy or to explain a failed build, and post agents to watch Production.
PIES AI
Switch the model
↓ nothing below changes ↓
PIES AI
Rules
By data class, role, application and region. Personal and card data are masked before the call. Seven presets, including Air-Gapped.
Memory
Organisation, app and agent notes in plain text, with author and time, stored in your own database. Read, edit or promote them on one page.
Record
Who asked, which policy and model, what was masked, tokens and cost, in a hash-chained, signed audit trail your security team can verify.
Memory, management and audit
This is what separates PIES Studio from an AI coding assistant. The AI is managed like any other enterprise system: it learns your standards, every call is routed by policy, and everything it does is written down.
PIES AI remembers how your organisation builds. Naming conventions, UI patterns, data rules and approved vocabulary are kept in Memory and applied to every build and agent run, whichever AI provider serves it.
A policy, not a model picker, decides which model may handle each request based on request type, data classification, role, application and region. Change the policy and every app follows it.
Every governed AI call is signed into a hash-chained audit log: who asked, which model answered, the data classes involved, the cost and the decision. Agents are attributed by name, run and version.
What you get
Most AI tools stop at a prototype. PIES Studio ships what an enterprise application actually needs.
Tables, relationships and an ER diagram, generated with no DDL. Connect existing MySQL, PostgreSQL, Snowflake or Zetaris data beside it.
Events, functions and a step catalogue, BPMN-style processes, PDF documents from templates, email and in-app notifications.
Password, Microsoft, Google, Apple, OpenID Connect and SAML set up once. Access groups enforced on the server, MFA, session policy, public screens.
Switch an application to mobile and the same model, functions and events render as a native Flutter app, previewed in a phone frame.
Deliveries change the app, errands act on it, posts stand watch on a schedule or webhook. Supervised or autonomous, with budgets and approvals.
One flag gives end users a themed assistant over their own data. Every generated app also exposes an MCP server and a REST API with Swagger.
Routing presets from air-gapped to frontier quality, redaction before a model sees data, a Test Console, and a signed hash-chained audit log.
PIES LLM on your own GPUs, a rented pod, a VM in your tenancy or an Apple Silicon Mac. Licence validated locally, weights delivered on media for air gaps.
Attach a BRS or specification and PIES AI builds from it. Organisation-wide knowledge reaches every build and agent run.
Branches and merge, rewind any AI turn, an activity log with restore, import and export, publish as a template, push code to GitHub or Bitbucket.
Preview, Test and Production per app. Deploy to Docker, Kubernetes, AWS or Azure with a control panel, rollback and start and stop schedules.
React or Angular front ends, Go, Python or Java back ends, Flutter for mobile, Docker packaging. Download it or run it without PIES.
Start from a finished application
Telco network operations, retail banking, B2B commerce with a storefront, hospitality, trades, professional services, education and more. Every template installs with sample data, a bundled agent pack, its own theme and the in-app chatbot switched on. Install, look around, then make it yours.
Editions
PIES Private AI on your own machine, nothing to pay. The full desktop IDE for building your products with AI that runs on your computer.
Self-hosted for your whole team, licensed per CPU, not per user.
PIES Studio hosted and run by us. Nothing to install.
No. You install PIES Studio yourself: Enterprise on your own servers or cloud account under a licence you buy, including air-gapped networks, or Hobby on your desktop. PIES never hosts your applications or data.
On Hobby, PIES Private AI Micro running on your own machine, free. Enterprise and PIES Cloud add Anthropic Claude, OpenAI, Azure OpenAI and xAI under your policies, plus PIES LLM private models included with every Enterprise licence and run on your own GPUs. A policy, not a picker, decides which model may see which data.
No. PIES Private AI Micro runs entirely on your computer and makes no network calls. We test that on every release. Nothing you type or build leaves your machine unless you connect a cloud model yourself.
Yes. The generated application is standard, portable source code in mainstream languages. You can export it and run it without PIES.
Per CPU on the server that runs PIES Studio, not per user. Unlimited developers on every plan, billed annually. Try it free for 60 days first.
Every AI request is recorded with user, model, data classification, cost and decision, in a hash-chained audit log with export.
Hobby is free on your desktop, no server needed. Enterprise runs on your own servers, on-premises, in your cloud or air-gapped, with unlimited developers, users and agents. 60 days free, no card, no sales call.
