> ## Documentation Index
> Fetch the complete documentation index at: https://docs.boostgpt.co/llms.txt
> Use this file to discover all available pages before exploring further.
# Workspaces
> Your agent's sandbox file system
## What are Workspaces?
Workspaces give your agent a sandbox file system — a place where it can create, read, edit, search, and delete files. When your agent uses **Agent** reasoning mode, it automatically has access to file tools that let it write code, generate documents, and manage project files.
**Workspace vs Training Data**: Training data teaches your agent *what it knows*. Workspaces give your agent *a place to work* — creating and editing files like a developer would.
## How It Works
Workspaces are automatically created when needed:
* **For logged-in users** — each user gets their own workspace per agent, so files are private
* **For anonymous visitors** — each chat conversation gets its own isolated workspace
No setup is required. When your agent is in **Agent** reasoning mode, the file tools are available automatically.
## What Your Agent Can Do
In Agent mode, your agent has access to these built-in file tools:
| Tool | What it does |
| ---------- | -------------------------------------------------- |
| **Write** | Create new files or overwrite existing ones |
| **Read** | Read file contents (can read specific line ranges) |
| **Edit** | Make precise, targeted changes to existing files |
| **Search** | Find patterns across all workspace files |
| **Delete** | Remove files from the workspace |
## File Explorer
When your agent creates or edits files, they appear in the **Workspace** panel on the chat page:
* **File tree** — browse all files organized by folder structure
* **Syntax highlighting** — code files are displayed with proper syntax coloring
* **Real-time updates** — files appear or update instantly when the agent makes changes
* **Manual editing** — you can edit file contents directly in the viewer
## Managing Workspaces
From the chat page, you can:
* **View files** — click any file in the workspace panel to view its contents
* **Create files** — add files manually via the workspace panel
* **Rename/move** — right-click files to rename or move them
* **Delete files** — remove files you no longer need
* **Switch workspaces** — if you have multiple workspaces, switch between them
## Best Practices
* Use Agent reasoning mode for tasks that involve creating or editing files
* Be specific in your instructions — "Create an HTML login form with email and password fields" works better than "make a form"
* Your agent can work across multiple files in a single turn — ask for a complete project structure
* Files persist across conversations, so you can return later to continue working
## Next Steps
Schedule your agent to run autonomously
Connect tools for your agent to use