Workspace

Workspace Context for AI Agents

Explore how Context is used within the Origin workspace to provide shared knowledge for AI agents. Learn to add, manage, and leverage context documents to improve agent output and task accuracy.

The Context panel in the workspace controls what shared knowledge agents use during a session. It is separate from the chat input, rather than sending information as a message, you attach it as persistent context that agents can draw from across the entire session.

The Context Panel

The Context panel is accessible from the right panel icon bar (document icon). The header shows the total number of project documents available and an Add Context button.

The panel is divided into two sections:

Project Documents

Lists the documents indexed in the project from Project Context. Each document shows its name and type (for example, Product Requirements or Documentation).

Two controls are available per document:

  • + Add to Context: adds the document to the active session's context so the agent uses it when reasoning during this session
  • The view icon (book icon), opens the document to inspect its contents

Documents listed here exist at the project level. Adding them to context makes them active for the current session specifically.

Recent Chats

Shows previous sessions from the project with timestamps and session type labels, for example, Worktree #1, Code Mode, or AI Task Discovery. Click + Add to pull a past session into the current context.

This is useful when continuing work that was started in a previous session, or when you want the agent to have awareness of prior decisions and reasoning without restating everything.

Adding Context

Click Add Context to open the Add Context modal.

Two methods are available:

Upload File: drag and drop one or more files into the upload area, or click to browse. Multiple files can be selected at once. Click Upload to add them to the project context.

Create Document: enter a Document Name and Content to create a new document directly. Both fields are required. Click Upload to save it.

Documents added here become available in the Project Documents list and can be added to any session's context.

Attaching Files Inline with @

You can also attach repository files directly in the chat input without opening the context panel.

Type @ in the message input to open the file and folder picker. It shows the full repository file tree with a search bar at the top. Use arrow keys to navigate, Enter to select a file, and Esc to close.

Selected files are attached to that specific message as context. This is useful for pointing the agent at a specific file for a focused question or change, rather than adding it as persistent session context.

Running Shell Commands with !

Type ! followed by a command in the message input to run it directly in the terminal without switching panels. This is a quick way to check output, run a script, or inspect the environment mid-conversation.

How Context Is Used

Once documents or files are added to context, agents draw from them automatically when:

  • answering questions about the system
  • generating or modifying code
  • proposing changes
  • reviewing work for alignment with requirements

You do not need to restate document contents in every message. Agents reference the available context as part of their reasoning.

Context vs Project Context Settings

The Context panel in the workspace controls what is active for the current session. The Project Context settings page is where documents are uploaded, indexed, and managed at the project level. Documents must exist in Project Context before they can be added to a session from the workspace panel.

On this page