Docs are external documents you upload to a workspace to give the agent additional context. They're separate from the project files the agent creates and manages. Think of them as background material that helps the agent understand your architecture, standards, compliance requirements, or any other reference material relevant to the project.
Drag and drop files into the Docs page of your workspace, or use the upload area to browse for files. Supported file types include PDF, Word documents, Excel spreadsheets, CSV, plain text, Markdown, JSON, XML, YAML, images, and common code files like Python, JavaScript, and TypeScript.
Each file can be up to 100 MB. If you upload a file with identical content to one that already exists, the system detects the duplicate and lets you know.
You can attach documents directly in the chat using the + button in the input area. Once a document has been uploaded to the workspace, you can also reference it in any future conversation by typing @ and selecting the document by name. Either way, the agent reads the content and uses it as context for its response.
Well-written docs directly improve the quality of the agent's output. The best use of workspace docs is project-specific context that helps the agent understand what you're building. Things like functional and nonfunctional requirements, technical architecture plans, project-specific security requirements, and design specifications.
For general standards and guidelines that should apply consistently, like compliance frameworks, naming conventions, or cloud architecture standards, rulesets are a better fit. Rulesets are enforced as rules the agent must follow. Docs are reference material the agent uses to understand the specifics of your project.
You can tag documents for organization. Tags are color-coded and scoped to the workspace. Use them to categorize documents by type, project area, or however makes sense for your team. You can filter the document list by tag to quickly find what you need.