Skip to content

Getting started with Flow 5

What is Flow?

Flow is an image search engine that uses AI models and is built on Apache Solr. It can handle millions of images on a single node and is designed to integrate with other platforms as a back-end system. Flow is commonly used to discover relevant image content, detect duplicate images, or find similar alternatives in large datasets.

The key features are:

  • Image search engine: Flow is an image search engine that allows users to search for relevant image content using various filters.
  • AI models: Flow uses fine-tuned AI models to maximize the relevance of search results.
  • Image content discovery: Discover relevant image content.
  • Duplicate detection: Detect exact and near-duplicates for copy-detection and content moderation tasks.
  • Similar alternatives: Find similar alternatives to an image.
  • Scalability: Handle millions of images on a single node. Use distributed search on multiple nodes to handle hundreds of millions of images.
  • Filters: Combine AI search with faceting, keyword search, range queries, and spatial search to help users narrow down their searches.
  • Apache Solr: Flow is built on Apache Solr, a popular open-source search platform.
  • Self-hosted: Own your data.

How to start

  1. Start the Flow Docker image and experiment with the provided sample dataset.
  2. Index your own data into Flow.
  3. Familiarize yourself with the advanced API parameters to get the full performance.