Skip to content

Ollama-OCR

Extract text from images with vision models through Ollama

Visit official site

What Ollama-OCR is

The tool is described as an OCR utility that uses recent vision language models through Ollama to extract text from images. The material states it ships as a Python package and a Streamlit web app.

What you can do with it

  • Choose vision models such as LLaVA 7B and Llama 3.2 Vision
  • Output Markdown, plain text, JSON, structured data or key-value pairs
  • Run batch processing in parallel and track each image
  • Resize and normalise images before recognition

Application scenarios

  • A researcher pulls data from academic paper images for analysis
  • A company digitises customer documents for storage and retrieval
  • A developer adds image-to-text conversion to an application

Getting started

  • Install the package with pip install ollama-ocr
  • Pull a vision model with ollama pull llama3.2-vision:11b
  • Import OCRProcessor and create an instance with a model name
  • Call process_image with the image path and the output format

Pros & cons

✓ What we like

  • Image resizing and normalisation are included as preprocessing
  • The Streamlit app supports drag-and-drop upload and text download

! What to watch out for

  • The package install and a pulled vision model are required before processing
  • Each process_image call needs an image path and an output format

FAQ

Which output formats are listed?

Markdown, plain text, JSON, structured data and key-value pairs.

How is a single image processed?

The card shows importing OCRProcessor, setting a model name, and calling process_image with a path and format.

Last reviewed: 2026-11-15

More AI image generation tools

View all →

How we review