# FileStack MCP for AI Agents AI Agent Connect

> FileStack MCP lets your AI agent handle heavy lifting for image and video assets. It can grab files from public URLs, run OCR on documents, tag objects in photos, and start video transcoding jobs. Use it to automate your media pipeline without manual editing or switching between multiple tools.

## Overview
- **Category:** image-video
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_KGFK0vBlAyeq8lyImuHLANnNzsFyop3aMeVzgPza/ai-agent-connect
- **Tags:** file-upload, image-processing, ocr, video-transcoding, content-analysis

## Description

Managing media assets usually means jumping between different tools to resize images, check for inappropriate content, or convert video formats. This Connector changes that by letting your AI agent handle the entire lifecycle of a file through a single interface. Instead of manually downloading a file, running it through an OCR engine, and then uploading it to a storage bucket, you can just tell your agent to do it. You can point the agent at a URL, have it extract the text, and then generate a specific CDN link for a thumbnail. It handles the heavy lifting of video transcoding in the background while you stay focused on your actual work. Because Vinkius hosts this Connector, you can connect it to your favorite AI client and get these capabilities instantly. It's about moving from manual file manipulation to a system where your agent understands and processes media as easily as it handles text.

## Tools

### get_sfw_status
Detects unsafe content in images. This helps automate community safety checks and moderation.

### generate_transform_url
Creates a CDN URL for resizing or filtering images. It lets you get specific dimensions without editing files.

### get_image_tags
Detects objects and features in an image. It is useful for auto-tagging a large photo library.

### get_metadata
Retrieves technical details like dimensions and mime types. Use this to check file properties quickly.

### get_ocr
Extracts printed or handwritten text from documents. It turns images into searchable text data.

### upload_from_url
Uploads a file to Filestack from a public URL. It is a fast way to move assets into your storage.

### get_video_status
Checks the status of a video transcoding job. Use this to see if your video is ready for use.

### start_video_transcode
Starts an asynchronous video transcoding job. It converts files into web-ready formats like MP4.

## Prompt Examples

**Prompt:** 
```
I have a link to a scanned receipt. Can you get the text from it?
```

**Response:** 
```
I've processed the receipt at that URL. Here is the extracted text:

| Date | Vendor | Total |
| :--- | :--- | :--- |
| 2023-10-12 | Coffee House | $15.50 |

I've also saved the metadata for your records.
```

**Prompt:** 
```
Are there any bad images in this folder of handles?
```

**Response:** 
```
I checked the following handles using the safety filter:

*   **Flagged:** 3 images (unsafe content detected)
*   **Cleared:** 47 images

Would you like me to list the flagged ones for review?
```

**Prompt:** 
```
Make a thumbnail for this video: https://example.com/video.mp4
```

**Response:** 
```
I've started a transcoding job for that video to create a web-ready version.

Here is your thumbnail URL:
https://cdn.filestack.com/handle/resize=w:300,h:200
```

## Capabilities

### Pull files from public URLs
The agent grabs files from the web and moves them into your Filestack storage automatically.

### Extract text from images
It turns printed or handwritten text in photos into actual text your agent can read.

### Identify objects in photos
The agent can detect specific features and objects within an image to help you organize assets.

### Check for unsafe content
It automatically flags images that contain unsafe or inappropriate material.

### Create optimized CDN links
You can generate specific URLs for resized or filtered images without editing the original files.

### Start video transcoding
It triggers cloud-based jobs to convert videos into web-ready formats like MP4 or HLS.

### Retrieve file metadata
The agent can pull technical details like dimensions and file types for any asset in your library.

## Use Cases

### Automated Content Moderation
A moderator needs to scan a gallery. The agent uses get_sfw_status on every image and flags the unsafe ones for review.

### Scanning Receipts for Data Entry
A data engineer has a folder of scanned receipts. The agent uses get_ocr to pull the totals into a structured format.

### On-the-fly Thumbnail Generation
A web developer needs thumbnails for user uploads. The agent uses generate_transform_url to create 200x200 versions instantly.

### Multi-format Video Distribution
A user wants to watch a video on mobile. The agent triggers start_video_transcode to create an HLS version for the player.

## Benefits

- Automate image tagging with get_image_tags to organize large photo libraries without manual entry.
- Extract text from scanned PDFs or photos using get_ocr to turn images into usable data.
- Generate resized CDN links with generate_transform_url to serve optimized images to your users.
- Check for inappropriate content using get_sfw_status to keep your platform safe and moderated.
- Handle video conversions via start_video_transcode so your agent can manage multiple formats at once.
- Move files from the web to your storage using upload_from_url to simplify your ingestion pipeline.

## How It Works

The bottom line is you get a way to manage complex media pipelines through simple natural language.

1. Subscribe to the Connector and add your Filestack API key to your agent's configuration.
2. Provide your agent with a file URL or a specific task like converting a video to MP4.
3. The agent executes the command and returns the file handle, a new URL, or the job status.

## Frequently Asked Questions

**Can the FileStack MCP read text from photos?**
Yes, it uses OCR to extract printed or handwritten text from images, making it easy to turn photos into data your agent can use.

**How does FileStack MCP handle large videos?**
It starts an asynchronous transcoding job, which means it processes the video in the cloud and lets you check the status later.

**Can I use FileStack MCP to resize images?**
You can generate CDN URLs that automatically resize, blur, or filter images without needing to edit the original file.

**Does FileStack MCP help with content moderation?**
Yes, it can automatically check images for unsafe content, which is great for keeping your platform clean and safe.

**Can I upload files to Filestack using the Connector?**
You can provide a public URL to the agent, and it will handle the upload to your Filestack storage for you.

**What kind of metadata can FileStack MCP get?**
It retrieves technical details like the file's dimensions and mime type so you do not have to check them manually.

**How can I extract text from a scanned document or image?**
Use the `get_ocr` tool with the file handle. The AI will perform Optical Character Recognition and return the extracted text along with its coordinates.

**Is there a way to check if an uploaded image contains adult or unsafe content?**
Yes, the `get_sfw_status` tool uses Filestack Intelligence to analyze the image and return a safety score, helping you moderate content automatically.

**Can I resize an image without actually downloading and re-uploading it?**
Absolutely. Use `generate_transform_url` to create a specific CDN URL with tasks like `resize=width:400`. This allows the Filestack CDN to handle the transformation on the fly.