What is PandaOS
An AI-powered desktop workstation that works with your local files and connected services.
The idea
Most AI tools live in a browser tab, isolated from the files, services, and workflows you actually use. PandaOS is a native desktop application that sits where your work already happens: your local filesystem, your email, your databases, your deployment pipelines.
Instead of copying context into a chat window, you point PandaOS at a project folder and it reads the files directly. Instead of describing what you need done in another tool, you connect that tool and PandaOS acts on it. The result is an AI assistant that operates with the same context and access you have, rather than a disconnected conversation partner.
Core principles
- Local-first. PandaOS runs on your machine. Your files, project history, and preferences stay on disk, not in someone else's cloud. Nothing leaves your computer unless you explicitly connect an external service.
- Connected, not siloed. Gmail, Google Drive, Supabase, Vercel, Trello, and 20+ other apps plug in as first-class integrations. PandaOS doesn't just read from them; it can send emails, run database queries, trigger deployments, and more, all from a single conversation.
- Adapts to how you work. Rules teach PandaOS your conventions. Skills give it reusable capabilities. Agents handle multi-step workflows autonomously. The more you configure, the less you repeat yourself.
- Model-flexible. PandaOS is not locked to a single AI provider. Switch between models and harnesses freely, and your projects, rules, and connected apps carry over.
What it looks like in practice
A few real examples of what a single PandaOS conversation can do:
- Open a codebase, refactor a component, start the dev server, and preview the result, all without leaving the chat.
- Summarize the last week of emails from Gmail, then draft replies based on a document in Google Drive.
- Query a Supabase database, build a chart from the results as an Artifact, and share it.
- Read a folder of research PDFs, extract key findings, and generate a slide deck.
The common thread: PandaOS combines understanding (reading your files and data) with action (writing, sending, deploying) in one place.