## Overview
The GoogleGeminiEmbeddingsConnector is a custom transformer that provides an integration to Google AI's Gemini models, enabling the generation of multi-modal embeddings. This enables FME to pass text, documents, images, or video files to the Gemini model and retrieve high-dimensional vector representations (embeddings) in return.
These embeddings capture the semantic meaning of your data and are ideal for downstream tasks such as semantic search, similarity matching, clustering, or populating vector databases. You can provide a text-only prompt, pass FME raster geometry directly or upload supported files including PDFs to generate your embeddings.
## Notes
Your own API Key is required to use this transformer. It is available via [aistudio.google.com](https://aistudio.google.com/app/apikey); currently Google offers a generous free tier for API calls to these models!
More information about the models, including embedding dimensionality, task types, and limitations:[ai.google.dev/gemini-api/docs/embeddings](https://ai.google.dev/gemini-api/docs/embeddings)
This transformer accepts all the text and multi-modal file types currently supported by the Gemini embedding models (including images and video via the Gemini File API).
## Licensing & Support
**License**: No license restrictions
**Cost:** Free
**Publisher:** Avineon Tensing
### Support
If you have any questions regarding this transformer, please email[uk@avineon-tensing.com](mailto:uk@avineon-tensing.com).
### Avineon Tensing
Avineon Tensing is a partner of Safe Software and a Value Added Reseller, with FME Certified Professionals and Trainers in the Netherlands, UK, France, Belgium, US and India. Avineon Tensing provides bespoke training services, consultancy and technical support for both FME Form and FME Flow.
To find out more about our FME capability visit [avineon-tensing.com](https://www.avineon-tensing.com) or email[info@avineon-tensing.com](mailto:info@avineon-tensing.com).
Would you like to know more? Click here to find out more details!