Skip to main content

Peliqan

Peliqan Machines: deploy vibe coded apps

Recent videos

Peliqan data platform

All-in-one Data Platform

Built-in data warehouse, superior data activation capabilities, and AI-powered development assistance.

Vibe coding gets you a working app quickly. Running it somewhere real is the part that usually stalls, because a multi-file React project needs a server, a database, a login and a way to deploy. Peliqan Machines are virtual machines that host your vibe coded web apps on Peliqan, with that backend already in place.

How to deploy a vibe coded app on Peliqan

  1. 1First, create a Machine. In Peliqan, open Peliqan Machine in the left menu and click + New machine. Give it a name and a short description. Peliqan AI reads that description and scaffolds a boilerplate web app, so you start from working code rather than an empty folder.
  2. 2Next, connect your coding agent. When the Machine is provisioned, click Connect your agent. You install the Peliqan CLI, sign in with your Peliqan account, connect to the Machine, and download an instructions.md file for your agent. That file tells the agent how to deploy and where the skill files sit inside the generated project.
  3. 3Then vibe code the app where you normally work. VS Code, Claude Code, or any other IDE. Tell your agent to read and follow instructions.md, then ask it to deploy. It pushes your code to the Machine over a secure SSH connection using the Peliqan CLI.
  4. 4Finally, preview it, then publish to production. Preview gives you a URL you can open and share outside Peliqan, Code shows every deployed file, and the deployment logs show what went wrong when something does. When it is ready, click Publish to production to make it permanently available to your users.

Machines or Data Apps: which one you want

Peliqan runs two kinds of app, and they suit different jobs.

  • Data Apps are single-file Python scripts with a Streamlit interface, written in the Peliqan editor. They are the simple option, and they are the right one for automations, data visualisations, data APIs and custom MCP servers.
  • Vibe coded web apps are multi-file projects in any language and any framework, written in your own IDE and deployed to a Peliqan Machine. They take a little more work, and in return you can build any kind of app.

If you have only ever built Data Apps, the short version is this: Machines are what you move to when Streamlit stops being enough. For the low-code route, see building Data Apps with low-code Python.

Peliqan as the backend for every app you build

Your app does not need its own backend, because it can use Peliqan as one. That covers authentication and authorisation through Peliqan, a built-in Postgres database, your Peliqan data warehouse with your business data in it, and your Peliqan connectors for reading from and writing back to your ERP or CRM.

Several apps can share the same backend. So the second app you build already has the login, the tables and the permissions waiting for it, and the two apps see the same data. The wider picture of what that backend does is in vibe coding on Peliqan, and vibe coding in-house apps shows a finished one.

What you can build on a Machine

  • Internal apps your team currently runs in a spreadsheet: timesheets, approvals, data entry, planning
  • A customer-facing portal that reads live data from the systems you have already connected
  • A rich front end in React, or any other framework, on top of your own business data
  • An app your AI assistant can also drive, by adding a custom MCP server on the same backend

Where to read more

The full setup, including the CLI, the preview URL, deployment logs and restarting an environment, is documented in Peliqan Machines – deploy vibe coded apps.

Ready to deploy your first vibe coded app? Get started with Peliqan.

FAQs

A Data App is a single-file Python script with a Streamlit interface, written in the Peliqan editor and run on Peliqan’s Data App VMs. A Peliqan Machine is a VM that runs a multi-file project in any language and any framework, written in your own IDE. Data Apps are simpler to write and deploy; Machines let you build any kind of app.

Any framework. That is the point of a Machine: the project is a normal multi-file codebase, so React, Vue, Next.js or anything else works. Streamlit only applies to Data Apps.

It is a file you download when you connect your agent to a Machine. It sets out the sequence of steps the agent follows to deploy your code, and it points to the skill files inside the generated project. Tell your agent to read and follow it, and deployment becomes one instruction rather than a setup task.

Deploying puts the app on a preview URL you can open and share. To make it permanent, click Publish to production in the top right. Preview is for iterating while you build; production is the version your users keep.

Get instant access to all your company data

Connect 300+ sources, serve any BI tool, and give every AI agent one governed endpoint to read, and write back where the app supports it.