> ## Documentation Index
> Fetch the complete documentation index at: https://docs.ospribrain.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Uploading Files

> Upload individual files, bulk ZIP archives, or import from Dropbox.

## Upload Methods

### Individual File Upload

<Steps>
  <Step title="Click + New, then Upload files">
    Or drag and drop files directly onto the main content area.
  </Step>

  <Step title="Select files">
    Choose one or more files from your computer.
  </Step>

  <Step title="Wait for processing">
    Files are uploaded to Supabase Storage (AES-256 encrypted), then automatically indexed for AI search. A brain icon appears once indexing completes.
  </Step>
</Steps>

<Frame caption="The + New dropdown menu with all upload options">
  <img src="https://mintlify.s3.us-west-1.amazonaws.com/ospribrain/images/vault-new-menu.png" alt="New menu" />
</Frame>

### Bulk ZIP Import

The Bulk ZIP Import is a 6-step AI-powered workflow for importing large document collections:

<Steps>
  <Step title="Upload">
    Drop a ZIP archive (max 500 MB, `.zip` format only) or click "Browse for ZIP file". The system automatically skips `__MACOSX` folders and `.DS_Store` files.
  </Step>

  <Step title="Parse">
    The ZIP is extracted and all entries are read. You see a progress indicator: "Reading archive..." then "Extracting entries..."
  </Step>

  <Step title="Analyze">
    The Brain AI analyzes the file structure, reading content patterns and generating folder suggestions. Progress shows: "Reading file structure...", "Analyzing content patterns...", "Generating folder suggestions...", "Finalizing..."
  </Step>

  <Step title="Review">
    The AI presents its organization plan with stats (total files, folders to create, files to commit). You can:

    * Review suggested folder assignments
    * Drag files between folders
    * Add context tags to folders
    * Move unassigned files
    * Read the AI's organization rationale
  </Step>

  <Step title="Commit">
    Click "Commit N files" to upload everything to your staging inbox. A progress bar shows upload status.
  </Step>

  <Step title="Done">
    Summary of created folders. Choose "Import another ZIP" or "Go to Staging Inbox" to review.
  </Step>
</Steps>

#### Context Tags (Bulk ZIP)

During the Review step, you can add tags to help organize and classify files:

**Domain tags:** Clinical, Regulatory, Research, Operations, Financial, Marketing, Legal, HR, Technical

**Team tags:** Medical Affairs, Operations, Compliance, Research, Executive

**Format tags:** Reports, Protocols, Data, Correspondence, Presentations, Contracts

**Sensitivity tags:** Restricted, Confidential, Internal, Public

Each sensitivity level has a distinct color: Restricted (red), Confidential (amber), Internal (blue), Public (green).

<Frame caption="Bulk ZIP import Review step with AI-suggested folder organization">
  <img src="https://mintlify.s3.us-west-1.amazonaws.com/ospribrain/images/vault-bulk-zip.png" alt="Bulk ZIP review" />
</Frame>

#### Supported File Types in ZIP

PDF, DOCX, DOC, PPTX, PPT, XLSX, XLS, CSV, TXT, MD, JSON, HTML, XML, PY, JS, TS, PNG, JPG, JPEG, GIF, MP4, MP3, and most common file formats.

### Import from Dropbox

See the dedicated [Dropbox Integration](/vault/dropbox) page for the full import and sync workflow.

### New Document

Click **+ New, then New document** to create a blank rich-text document directly in the Vault. A WYSIWYG editor opens with formatting toolbar (headings, bold, italic, lists, links). Content auto-saves.

### New Folder

Click **+ New, then New folder** to create a subfolder. Enter a name in the dialog and click Create.

## After Upload

Once uploaded, files go through this pipeline:

<Steps>
  <Step title="Stored">
    File is saved in Supabase Storage with AES-256 encryption at rest.
  </Step>

  <Step title="Text extracted">
    OCR for images, text parsing for documents, metadata extraction for all types.
  </Step>

  <Step title="AI Summary generated">
    A brief summary of the file's content is auto-generated and shown in the detail panel.
  </Step>

  <Step title="Vector embeddings created">
    Content is embedded using OpenAI's `text-embedding-3-large` model (1024 dimensions).
  </Step>

  <Step title="Indexed in Pinecone">
    Embeddings are stored in Pinecone under the vault's namespace, making the file searchable by the Brain AI.
  </Step>
</Steps>

<Info>
  Files in vaults with Brain Access turned **OFF** skip the embedding and indexing steps. They are stored but not searchable by AI.
</Info>