Dev Servers
Dev servers create a live feedback loop where Panda edits code and you immediately see the result in an embedded browser preview.
The live feedback loop
The most powerful thing about dev servers in PandaOS is the feedback loop they create. You ask Panda to change something. Panda edits the code. The dev server picks up the change via hot reload. The embedded browser preview updates. You see the result in seconds, right next to the chat, without switching windows or refreshing manually.
This tight loop changes how you work with AI. Instead of describing what you want, waiting, reading a diff, mentally simulating the result, and then running the app to check, you just watch it happen. If it is not right, you say so and iterate.
Automatic framework detection
When you open a web project, PandaOS detects the framework and suggests the right dev command. For a Next.js project, that might be npm run dev; for a Vite project, npx vite. You can accept the suggestion or customize it. The detected framework also informs Panda's understanding of your project, so it knows whether to generate App Router code or Pages Router code, for example.
Running multiple servers
Many real projects need more than one process. A frontend and an API server. A web app and a background worker. PandaOS supports multiple dev servers per project, each with its own command, port, working directory, and auto-start preference.
This is particularly useful in monorepos. You might run the Next.js frontend from apps/web/ on port 3000 and the Express API from apps/api/ on port 4000, both managed from the same project. The browser preview connects to whichever server you choose.
Browser preview integration
Once a dev server is running, the embedded browser preview automatically opens to its URL. The preview sits alongside the chat and updates as the dev server reflects code changes. This means Panda can make an edit, you can see the visual result, and you can respond with feedback, all without leaving PandaOS.
For the full set of workspace tools available alongside the preview, see Files, Terminal & Browser.
Environment variables and credentials
A dev server usually needs secrets, so PandaOS gives the project a Credentials panel: the .env files it finds, with their variables listed and values hidden until you reveal them. You can add, edit and delete variables there instead of opening the file by hand.
The agent reads and writes those variables through the same panel rather than by opening .env in the editor, so a key can be set for a run without its value being printed into the conversation. Credentials open as their own panel and can be placed as a card on the Canvas.
Knowledge
Knowledge files give Panda project-specific context that goes beyond the code itself, like specs, architecture docs, and domain reference material.
Git & AI Commits
PandaOS integrates Git directly into the workspace and uses AI to generate commit messages from real diffs, keeping your history clean without the effort.