Genexy takes your favorite AI agents to the next level with powerful MCP tools. They understand your project, build and connect modular components, and produce high-quality code while maintaining a clean, consistent architecture.
One environment
Any AI
Sign in with the subscription you already pay for, bring your own API key, or run models on your own hardware. Use several agents side by side in one project.
Native
Genexy stays fast and lightweight, even on massive projects. Save your RAM and CPU for builds, containers and AI agents, not your editor.
~1 s
Startup
From launch to a ready IDE
~50 MB
Core RAM
Plus ~400 MB for the system WebView
Measured on Intel Core Ultra 7 255U·32 GB RAM·Windows 11 Pro 25H2
Native Go core with the system webview. No Electron, no bundled Chromium.
Built onWailsThe core idea
Describe your components and the contracts between them, such as OpenAPI. Genexy generates the structure from those contracts, and through MCP your AI implements the logic inside it, so AI spends its effort only where reasoning is needed.
Clients
Gateway
Services
Databases
Messaging
clients → gateway · gateway → services · service → its own database · everything on the NATS bus
A preview of the visual Architecture Editor, which is in development. Today you describe the same components and contracts in the project specification and generate the backends from them.
For AI
Instead of seeing only files and folders, AI can work with the components, their parts and how they connect.
For long-lived projects
AI cannot edit the files a generator owns. To change them, it changes the contract and regenerates, so the structure stays consistent however much code AI writes.
Workflow preview
Scroll through a real Genexy workflow.
Nothing here is staged: this is a recorded run with the public Binance Spot OpenAPI. Give the same prompt to Genexy AI on your own machine and you get the same backend.
This demo is real and you can repeat it. Install Genexy, open an empty folder and send this prompt to Genexy AI:
Download the Binance Spot OpenAPI from github.com/binance/binance-api-swagger and build a Spring Boot backend in Kotlin from it. No business logic yet.Download Genexy
Desktop · Tablet · Phone
The same Genexy on a phone and a tablet, with its own project filesystem, Genexy Shell, Git and built-in AI integrations. Not a remote viewer, not a code preview.
Projects & files·Editor·Search·Git·Genexy Shell·AI integrations
Desktop capability
Local CLI agents such as Kiro, Claude Code or Codex CLI run on desktop. With Remote Access (experimental) a phone or tablet works directly in your desktop workspace, including its terminals and agents.
Community
Genexy is in public Beta. Join other developers testing new builds, sharing feedback and helping shape what comes next.
Join DiscordFeatures
Each feature links to its page in the documentation.
Plans
FREE
€0
A complete IDE with multi-agent AI on every platform. No account required.
PRO
€19/ month
The tools that let AI work efficiently: better, more maintainable projects for a fraction of the tokens.
Included in PRO
Licenses are issued on our Discord
Enterprise
Technology partnership or sponsorship? Write to us as well.
Behind Genexy
I started Genexy because I wanted architecture, AI, automation and everyday development to work as one system, not as a collection of disconnected tools. The cross-platform IDE is already available in Beta, and more of the architecture-first workflow ships as each part becomes ready.
Clear answers
No. The editor, terminal, Git and AI panel are how you work with Genexy, but they are not what Genexy is. Genexy is built around the architecture of your software: components, contracts such as OpenAPI, and the code generated from them. Every IDE capability is also open to your AI. Generating backends from contracts works today; the visual Architecture Editor is in development.
No. You decide the architecture and the contracts. Genexy generates the repetitive parts from them (project structure, models, controllers, service interfaces), and your AI implements the business logic in the stubs it gets. AI is not allowed to edit generator-owned files: to change them, it has to change the spec and regenerate.
Spring Boot in Java or Kotlin, Go, NestJS, Node.js (Express, Fastify or node:http), ASP.NET Core and Rust on axum, from OpenAPI contracts. When the contract changes, you regenerate and your implementation stays. Code generation is a PRO feature and still experimental.
Through Genexy MCP, your AI agent can use most of what you use in the IDE: files, search, Git, the shell, run configurations, the debugger, Docker and databases. You see every tool call live, approve actions one by one or switch on Autopilot.
For free: the editor with code intelligence for 25+ languages, search, Git, native terminals, Genexy Shell on every platform, run configurations, the debugger for Go, Java and Kotlin, and AI chat with Kiro, Claude Code, Codex, Copilot, Cursor or your own endpoint. PRO adds Genexy MCP, code generation, Docker, SSH and database connections and a dependency scan with vulnerability advisories.
Yes. Mobile Genexy runs inside the operating-system sandbox with its own project filesystem, Genexy Shell, Git and built-in tooling. Local CLI agents such as Kiro, Claude Code or Codex CLI run on desktop; with Remote Access (experimental) your phone or tablet works directly in the workspace of your desktop, including its terminals and agents.
Not by default. Local projects stay on your device unless you intentionally use a feature that sends selected content elsewhere, such as an AI provider or a support request. Remote Access connects your own devices and its session is encrypted. Genexy does not use your source code to train AI models.
PRO costs €19 / month. To get a license, message us on Discord.
