About fal-vision
fal-vision is a skill for analyzing images with fal.ai vision models. Its script supports segmentation, detection, OCR, captioning, and visual question answering, with an option to specify a model endpoint.
Key features
- 01Image object segmentation
- 02Object detection
- 03OCR text extraction
- 04Image captioning and visual QA
- 05Specify a model endpoint
Use cases
- 01Locate a specific object in an image
- 02Identify objects in pictures
- 03Extract text from document images
- 04Answer questions about image content
MCP support
The README mentions using the `search_models` MCP tool to find vision analysis models by keyword; the skill itself runs image analysis through a script.
Install
Copy the command and run it in a terminal.
npx skills add https:/After installing, type/fal-visionin Claude Code to confirm it loaded.
Bundled filesscripts/analyze.sh
Details
From the project READMEfal-vision
Analyze and understand images using fal.ai vision models — segmentation, detection, OCR, captioning, and visual QA.
Scripts
| Script | Purpose |
|---|---|
analyze.sh | Analyze an image (segment, detect, OCR, describe, QA) |
Usage
Segment Objects
./scripts/analyze.sh --image-url "https://example.com/photo.jpg" --operation segment --query "the red car"
Detect Objects
./scripts/analyze.sh --image-url "https://example.com/photo.jpg" --operation detect
Extract Text (OCR)
./scripts/analyze.sh --image-url "https://example.com/document.jpg" --operation ocr
Describe Image
./scripts/analyze.sh --image-url "https://example.com/photo.jpg" --operation describe
Visual QA
./scripts/analyze.sh --image-url "https://example.com/photo.jpg" --operation qa --query "How many people are in this image?"
Arguments
| Argument | Description | Required |
|---|---|---|
--image-url | URL of image to analyze | Yes |
--operation | segment, detect, ocr, describe, qa | Yes |
--query / -q | Text prompt for segment/qa operations | For segment/qa |
--model / -m | Override model endpoint | No |
Finding Models
To discover the best and latest vision/analysis models, use the search API:
# Search for segmentation models
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --query "segmentation"
# Search for object detection models
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --query "object detection"
# Search for OCR models
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --query "ocr"
# Search for image captioning / visual QA models
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --query "caption"
bash /mnt/skills/user/fal-generate/scripts/search-models.sh --query "visual question"
Or use the search_models MCP tool with keywords like "segmentation", "detection", "ocr", "caption", "vision".