Reverse Image Search
Build a reverse image search engine that accepts an image as input and returns visually similar results from your indexed collection. Uses multimodal embeddings to match visual features like color, shape, composition, and semantic content.
from mixpeek import Mixpeekclient = Mixpeek(api_key="YOUR_API_KEY")# Create a collection with image embeddingscollection = client.collections.create(namespace_id="ns_your_namespace",name="product_images",extractors=["multimodal-extractor"])# Upload your image catalogclient.buckets.upload(bucket_id="bkt_your_bucket",url="s3://your-bucket/images/")# Reverse image search: query with an image URLresults = client.retrievers.execute(retriever_id="ret_your_retriever",query={"image_url": "https://example.com/query-photo.jpg"})for doc in results["results"]:print(f"Match: {doc['document_id']} (score: {doc['score']:.3f})")
Feature Extractors
Retriever Stages
mmr
Reorder results using Maximal Marginal Relevance for diversity
