Generate Alternate Text BLIP

Runs locally and uses the BLIP large model to describe image content in alternative text. For PDF output without watermarks, a PDFix SDK license is required.

Table of Contents

• Generate Alternate Text BLIP
  • Getting started
  • Usage
  • Commands
  • Arguments
  • Examples
  • Model
  • Help & support
  • Licenses

Getting started

You need Docker installed. The first run downloads the image and may take longer than later runs.

Usage

Mount a folder into the container and run a subcommand:

docker run --rm -v "$(pwd)":/data -w /data pdfix/alt-text-blip-large:latest <command> [options]

Commands

• generate-alt-text: Generate alternate text (PDF → PDF or supported image → TXT)

Arguments

generate-alt-text

Option	Required	Type / expected value	Description
--input, -i	yes	Path to an existing .pdf or supported image file	Input PDF or image
--output, -o	yes	Path for output .pdf or .txt (must match mode)	Output file
--model	no	Path to model directory inside the container (default: model)	Local BLIP model path
--overwrite	no	Boolean string: true/false, yes/no, 1/0 (default: false)	Overwrite existing Alt text
--zoom	no	Float (default 2.0)	Page render zoom for PDF mode
--name	no	String (PDFix account license name)	PDFix license name
--key	no	String (PDFix account license key)	PDFix license key

Examples

Generate alternate text for figures in a PDF:

docker run --rm -v "$(pwd)":/data -w /data pdfix/alt-text-blip-large:latest \
  generate-alt-text --name "${LICENSE_NAME}" --key "${LICENSE_KEY}" \
  -i /data/input.pdf -o /data/output.pdf --model /model

Caption a single image to a TXT file:

docker run --rm -v "$(pwd)":/data -w /data pdfix/alt-text-blip-large:latest \
  generate-alt-text -i /data/image.jpg -o /data/output.txt --model /model

Model

The image includes the BLIP large model and runs fully offline (CPU).

Help & support

For PDFix SDK licensing or issues, contact support@pdfix.net.

Licenses

• PDFix Terms
• BLIP large model — BSD-3-Clause