◬
Get Started

Files & Storage

Magus lets you attach files to conversations so your agents can read and reference them. You can also access files from previous conversations, share files within a folder or workspace, and manage your overall storage from account settings.

Uploading Files

In Chat

The easiest way to upload a file is directly in a conversation:

  • Drag and drop: Drag a file from your desktop into the chat window and drop it anywhere in the message area
  • File picker: Click the attachment icon (paperclip) in the message input area and select a file from your device

The file will be attached to your message. Once the message is sent, the agent can access the file and reference its contents in its response.

From the File Browser

For files you want to use across multiple conversations, open the File Browser from the workbench’s Files icon, or visit /files directly. From the top bar of the browser:

  • Click Upload to pick files from your device. They upload into the current view (the open folder, or your root files if none is open).
  • Drop files anywhere on the grid to upload them in place.

Files uploaded here are available immediately and persist beyond any single conversation.

The File Browser

The file browser is a Drive-style view for everything you have uploaded or that has been generated for you. Click the Files icon in the workbench sidebar to open it.

Sidebar entry points

The sidebar lists places to start from:

  • My Files: your personal files and folders.
  • Shared with me: when you are in a team workspace, files others have shared with the team.
  • Recent: files modified in the last 30 days.
  • Templates: files you have marked as templates so they can be reused.
  • Connected Sources: collections synced from external services. Click the chevron to see them.
  • Trash: files you have moved to trash. Read-only in this version.

Below the entry points are three filter pills: Type, Modified, Source. Pick a value to narrow the current view, or pick “Any” to clear.

The bottom of the sidebar shows your storage usage at a glance.

Top bar

The browser’s top bar has:

  • Breadcrumbs: click any segment to jump back to a parent folder or scope.
  • Search: filter the current view by name. The URL updates as you type, so you can bookmark a search.
  • Sort: sort by Modified (default), Name, or Size, ascending or descending.
  • List / Grid toggle: switch between table and card views. Your choice is saved on your account.
  • + New folder: insert a new folder in the current location.
  • Upload: pick files from the file system.

Folder navigation

Click a folder card or row to open it. The breadcrumb updates and the URL changes so you can bookmark or share a folder view.

File actions

Click a file to open its detail view in a new tab. Right-click any file or folder for the action menu:

  • Open, Open in new tab, Download
  • Open chat about this file
  • Rename, Move to, Share to workspace
  • Toggle template (files only)
  • Move to trash

Selecting Move to opens a folder picker so you can choose the destination. Selecting Rename opens an inline editor.

The trash view is read-only in this version. Restoring or permanently deleting files will be added in a future update.

File Scopes

Every file has a scope that controls who can access it and in which conversations.

Scope Who can access it
Chat Only the current conversation
Folder All conversations inside the same folder
Workspace All members of your workspace (team access)
Global All your conversations, across all agents

Choose the right scope based on how broadly you need the file to be available. For sensitive documents, Chat scope keeps the file isolated to one conversation. For reference material your whole team uses, Workspace scope makes it available everywhere.

Downloading Files

To download a file:

  • In a conversation, click the file attachment to expand it, then click the Download button
  • In the Files tab, find the file and click the Download icon next to it

Agent-generated files (such as images generated by an image model or outputs from code execution) appear in the conversation and can be downloaded the same way.

Supported File Types

Magus can extract text from a wide range of file formats:

  • Documents: PDF, Word (.doc, .docx), RTF, EPUB, OpenDocument (.odt)
  • Spreadsheets: Excel (.xls, .xlsx), CSV, OpenDocument (.ods)
  • Presentations: PowerPoint (.ppt, .pptx)
  • Images: JPEG, PNG, GIF, WebP, TIFF, BMP, SVG (text extraction via OCR)
  • Web and data formats: HTML, XML, JSON, YAML, Markdown
  • Plain text: .txt and other text-based files

Large files are processed in chunks for semantic search. After uploading, there may be a brief processing period before the file is fully available to your agent.

Storage Limits

Your available storage depends on your subscription plan:

  • Free plan: Limited storage included
  • Paid plans: More storage, with higher per-file upload limits

You can see your current storage usage in Account Settings under Storage. The display shows how much you’ve used and how much your plan allows.

If you reach your storage limit, you’ll need to delete files or upgrade your plan before uploading more.

Managing Storage

Viewing Storage Usage

The bottom of the file browser sidebar shows a quick storage bar with your current usage. For a fuller breakdown, open Account Settings and look for the Storage section.

Deleting Files to Free Up Space

To delete a file:

  1. Open the file browser at /files
  2. Find the file (use Search or a filter to narrow down)
  3. Right-click the file and choose Move to trash

The file moves to the Trash entry point and is hidden from your active views. It still counts against storage in this version. Permanent deletion to fully free space will be added in a future update.

Workspace Files

If you’re on a team workspace, files with Workspace scope are shared among all members. Only the person who uploaded a file (or a workspace admin) can delete it.

File Processing Status

After uploading, files go through a processing step where Magus extracts text, creates chunks for semantic search, and stores the content. Open a file from the browser to see its status:

  • Pending: Upload received, processing hasn’t started yet
  • Processing: Being indexed for semantic search
  • Ready: Fully available to your agent
  • Error: Processing failed. Try deleting and re-uploading the file.