Media & Storage

Upload documents, images, and videos that your AI agents can access through built-in MCP tools. Give your agents context from your own files.

Overview

MCP Gateway Pro includes a built-in media server that exposes your uploaded files as MCP tools. When you upload a document, image, or video through the console, any AI agent connected to your gateway can list and retrieve those files.

Text content is automatically extracted from PDFs and text files, so agents can read document contents directly without needing a separate parser.

Storage: Included in the AppXen Platform appliance — files are stored on your own S3 bucket, billed on your normal AWS bill.

1

Upload Files

Navigate to MCP Gateway → Storage in the console. Drag and drop files onto the upload area, or click to browse.

You can upload multiple files at once. Each file can be up to 500 MB.

Supported file types

Documents & Code

.pdf .txt .md .json .yaml .csv .py .js .ts .html .sql .go .rs .java and more

Images & Video

.png .jpg .gif .webp .mp4 .mov .webm

2

Access from AI Agents

Uploaded files are automatically available to any AI agent connected to your gateway through two built-in tools:

media__list_media

Returns a list of all uploaded files with their metadata (filename, size, type, upload date). 

{
  "method": "tools/call",
  "params": {
    "name": "media__list_media",
    "arguments": {}
  }
}

media__get_media

Retrieves the content of a specific file. The response format depends on the file type: 

{
  "method": "tools/call",
  "params": {
    "name": "media__get_media",
    "arguments": {
      "media_id": "your-media-id"
    }
  }
}
PDF

Returns extracted text content from all pages

Text

Returns the raw file contents (code, markdown, CSV, etc.)

Image

Returns base64-encoded image data for inline display

Video

Returns a temporary signed URL valid for 1 hour

3

Automatic Text Extraction

When you upload a PDF or text-based file, the gateway automatically extracts the text content and stores it alongside the original. This means your agents can read document contents directly — no additional parsing required.

Example: Upload a 50-page product spec PDF. Your agent calls media__get_media and receives the full text, ready to answer questions about the document.

Text extraction is supported for PDFs and all text-based formats (source code, markdown, CSV, JSON, YAML, etc.).

Storage & Billing

Your storage usage is shown in the console under MCP Gateway → Storage. Uploaded files are stored in your own S3 bucket, in your own AWS account, and AWS bills that storage to you on your normal AWS bill. AppXen does not meter or bill for storage. See Pricing for how the appliance is billed.

Managing Files

The Storage page in the console lets you view all uploaded files, check their type and size, and delete files you no longer need. Deleted files disappear at once from the console and from agent access. The bucket keeps a deleted file as a noncurrent version for 30 days, and then S3 removes it.

What's Next?