Skip to content

voyage-multimodal-3

A multimodal embedding model for text and screenshots

Visit official site

What voyage-multimodal-3 is

The model is described as a multimodal embedding model from Voyage AI that turns text and content-rich images into vectors for document retrieval.

What you can do with it

  • Retrieve legal documents from clause screenshots
  • Search financial reports containing tables
  • Find academic documents with teaching charts

Key details

  • Visual cues include font size and text position
  • A unified representation handles both modalities
  • The material states a 19.63 percent average gain

Pros & cons

✓ What we like

  • Knowledge bases can be vectorised directly
  • A free trial and API access are described

! What to watch out for

  • Registration and API access are part of the steps
  • Integration into a knowledge system is left to the user

FAQ

What does it vectorise?

The material names text and screenshots of PDFs, slides and charts.

Is there a reported figure?

The material states an average 19.63 percent gain in retrieval accuracy.

Last reviewed: 2026-11-15

More AI search engine tools

View all →

How we review