Image Captioner - Alt Text & Photo Descriptions

AI Image Captioner

Generate accurate, detailed text descriptions for any image using AI - bulk caption product photos, screenshots, or any image URL for SEO alt text, accessibility compliance, and content tagging.

What you get

• Natural-language captions generated by Molmo 2, trained on 712,000+ human-described images
• Three detail levels: brief one-liner, balanced description, or full detailed paragraph
• Optional focus directive to target specific aspects (text, background, faces, objects, etc.)
• One output record per image with the caption and source URL
• Supports bulk processing - pass up to 50 image URLs and get all captions in a single run
• Export to JSON or CSV directly from the Apify console

Use cases

• E-commerce SEO - generate alt text for thousands of product images automatically
• Accessibility compliance - add descriptive alt text to images on websites and apps
• Content moderation - understand what's in user-uploaded images before publishing
• Dataset labeling - annotate image datasets for machine learning pipelines
• Digital asset management - auto-tag and describe photos in large media libraries
• Social media monitoring - caption scraped images to make them searchable by content

Examples

Image	Detail Level	Caption
	High	A vibrant red Nike sneaker takes center stage in this striking advertisement, set against a bold red background that creates a visually cohesive and eye-catching composition. The shoe is positioned at an angle, giving the impression of motion and energy. The sneaker features a white Nike swoosh, darker red laces, and "Nike Free" branding on the white sole. The lighting is bright and even, highlighting the shoe's textures and details.
	Medium	A vibrant red Nike sneaker is displayed against a matching red background. The shoe features a white Nike swoosh and "Nike Free" branding on the sole. The laces are a darker shade of red, complementing the overall design.
	Low	A red Nike sneaker with white accents.
	High	A top-down view of a rustic wooden table with three round bowls arranged in a triangular formation. The central bowl features slices of medium-rare steak garnished with fresh green leaves and a red chili pepper. The left bowl holds crispy fried fish topped with a creamy sauce and herbs. The right bowl contains a meat dish garnished with thinly sliced red onions and nuts. Scattered around the bowls are whole chili peppers, cashews, and a small bowl of brown dipping sauce.
	Medium	Three bowls of food are arranged on a gray wooden table, creating a rustic dining scene. The central bowl contains sliced steak, while the left bowl holds fried fish topped with sauce and herbs. The right bowl features a meat dish garnished with onions and nuts.
	Low	Three bowls of food on a wooden table with garnishes.

How to use

1. Paste one or more image URLs into the Images field (or upload files directly)
2. Choose a Detail Level - High gives the most descriptive output (recommended for SEO and accessibility)
3. Optionally add a Focus hint to direct the model's attention (e.g. "describe only the text visible")
4. Click Run - captions appear in the Dataset tab when complete
5. Export results as JSON or CSV, or connect to downstream actors via the Apify API

Output format

Each dataset record:

{
  "inputImageUrl": "https://example.com/product.jpg",
  "caption": "A white ceramic coffee mug sitting on a wooden table next to an open laptop. The mug has a minimalist logo on the front and steam rising from the top, suggesting the coffee is hot.",
  "detailLevel": "high",
  "status": "success",
  "error": null
}

Input options

Field	Type	Description
Images	URL list	One or more http/https image URLs or base64 data URIs
Upload Images	File upload	Upload images directly from your computer
Detail Level	Select	Low (one-liner), Medium (balanced), High (detailed paragraph) - default: High
Focus	Text	Optional directive to focus the caption on a specific aspect of the image

Limits

• Maximum 50 images per run
• Each image must be a publicly accessible URL or a base64 data URI
• Processing time is typically 5-15 seconds per image

Related AI image actors

Part of a complete AI image toolkit - explore the rest of the suite:

• AI Image Background Remover - Remove backgrounds to clean transparent PNGs
• AI Image Upscaler - Batch-upscale images to 4K or 8K
• AI Image Watermark Remover - Remove text and logo watermarks from images
• Image OCR Scraper - Extract text from images in 109 languages
• Photo Location Finder - Find where a photo was taken - no EXIF needed