.svg)
A Workspace is an isolated project environment that gives you complete control over code, data, documentation, and AI models. Each Workspace maintains strict boundaries by default - models can only access data within their Workspace unless you explicitly configure external connections. Your code runs in private compute environments, your data stays secure, and your AI assistants operate within defined boundaries.
Think of it as your own digital laboratory where computation, data, and AI work together continuously. You set up the tools and data you need, then let your automated processes run securely in the cloud. As results come in, AI assistants can analyze them, update documentation, and alert you to important findings - all while maintaining isolation from other projects and users.
You can create multiple Workspaces to support different projects or to isolate materials for different teams.
Your Workspace's file system provides secure storage for code, data, and documentation. Files are organized in a hierarchical structure and are only accessible to the models and team members you authorize. Read more about Files.
Each Workspace includes its own dedicated compute environment. This private runtime executes your code in isolation, ensuring that computations remain separate from other projects and users. Whether running standalone scripts or embedded computations within notes, all code executes in this secure environment. In future releases, these compute tasks will also be able to run continuously in the background. Read more about Compute.
Notes serve as a shared canvas between you, your team, and your AI models. They support rich text, mathematical notation, and embedded computations. Workers can automatically update these notes as part of their execution, maintaining a living knowledge base that evolves with your project. Notes can reference other Workspace components, creating a connected source of truth for your team. Read more about Notes.
The chat system enables direct interaction with AI models while maintaining context awareness. Models can reference your Workspace's files, notes, and computation results during conversations, but cannot access anything outside the Workspace boundaries. Read more about Chat.
Workers are AI models that you deploy within your Workspace. Each Worker operates within the Workspace's boundaries, able to access the files, notes, and computations contained within. When you choose, Workers can also respond to external triggers - updating notes when builds complete, documenting changes when pull requests merge, or analyzing new data as it arrives. Read more about Workers.
Think of your Workspace as a system that's always running, ready to help your team work more effectively.
You might set up:
The key is that these processes run continuously in your Workspace, maintaining their context and keeping your team's knowledge base current and accurate.
Further, Workspaces are collaborative environments. You can invite other users such as team members, contractors, and external collaborators to join your Workspace -- allowing you to work together on defining and automating your workflows.
