- RAW
- PARSED
- SAVED
- REVIEW
Short answer
How can I use my AI to process documents in Azure?
Your AI uses the Azure Blob Container MCP to pull raw files from your storage, parse the content, and write the structured data back to a new blob. You get a direct path from messy files to organized data.
Where your data lands.
The outcomes of automation.
The AI moves data through these stages based on your instructions.
RAW
Unstructured files
The AI identifies and pulls the original documents from your container. These stay exactly as they were uploaded.
PARSED
Structured data
Your agent extracts specific fields or summaries from the files. You get clean, usable information instead of raw text.
SAVED
New storage blobs
The AI writes the processed results back to Azure. This keeps your organized data right next to your source files.
REVIEW
Human verification
You decide if the extracted data is correct. The AI won't delete or overwrite anything without your direct command.
The workflow
What your AI does when the files arrive.
The AI acts as the bridge between your raw storage and your structured data.
Scan the container
The AI looks through your storage to find the specific files or folders you want to process.
list_blobsRead the content
Your agent pulls the actual data from the identified files so it can read and understand the text.
get_blobExtract information
The AI analyzes the file content to find the specific details you asked for, like dates, totals, or names.
get_blobWrite the results
The AI saves the newly structured information back into a new file in your Azure container.
put_blob
Try it
Copy these to start.
Use these prompts to kick off your first automation task.
Starting points
These are opening lines. Swap in your specific container names and file types.
Azure Blob Container Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_KegC9wyMjVWOSEhCaSwb9DkXf1hOZWEJE3EaqGTJ/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Azure Blob Container capabilities are ready to use.
{
"mcpServers": {
"azure-blob-container-mcp": {
"url": "https://edge.vinkius.com/vk_preview_KegC9wyMjVWOSEhCaSwb9DkXf1hOZWEJE3EaqGTJ/mcp"
}
}
}List all the blobs in my 'invoices' container so I can see what needs processing.
Read the file 'report_2023.pdf' from my storage and summarize the key findings.
Find all CSV files in the 'data-drop' container and extract the total sales from each.
Take the summary you just created and save it as a new JSON file in my 'processed' container.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Azure Blob Container once, then ask.
You only need to link your Azure account once. Your credentials stay encrypted, and you can immediately start giving your AI instructions through any compatible client.
Connect Azure Blob Container to your AIFAQ
How this task behaves
- 01
Can the AI delete files from my Azure container?
No. The current tools only allow the AI to list, read, and write blobs. It cannot delete anything.
- 02
Will the AI overwrite my existing files?
The AI can only overwrite a file if you specifically tell it to save a new blob with the exact same name as an existing one.
- 03
Does the AI need to see my Azure password?
No. Your credentials are encrypted and handled securely by Vinkius; your AI client only uses the connection to perform the tasks.
- 04
Can the AI process images or just text files?
The AI can process any file type that your AI client is capable of reading, such as PDFs, text files, or images, once the blob is retrieved.
- 05
Can the AI create new containers for me?
No. The AI can only interact with blobs within containers that already exist.
More questions about Azure Blob Container? The Connector page answers them. See everything the Azure Blob Container Connector can do
Connect Azure Blob Container to Claude, Cursor, ChatGPT & more
