Amazon S3 Vectors introduces metadata pre-filtering for up to 5x higher recall on filtered search

Amazon S3 Vectors Updates

Amazon S3 Vectors now supports pre-filtering, which evaluates metadata filters before running similarity search, returning up to 5x more of the matching vectors when your filter is selective. S3 Vectors also adds a prefix match operator ($startsWith) for filtering on values like paths and URLs. These improvements enhance retrieval-augmented generation (RAG), agentic, and semantic-search applications by providing more complete results when filtering, leading to more relevant answers.

Indexes in new vector buckets use metadata pre-filtering by default. To use pre-filtering on an existing index, update it in place with the UpdateIndexMode API. You can compare pre-filtering against your current filtering on the same index using a per-query parameter on QueryVectors.

What to do

• Update existing indexes to use metadata pre-filtering with the UpdateIndexMode API.
• Compare pre-filtering against current filtering using the per-query parameter on QueryVectors.
• Use the AWS CLI, AWS SDKs, or the Amazon S3 console to get started.

Source: AWS release notes

If you need further guidance on AWS, our experts are available at AWS@westloop.io. You may also reach us by submitting the Contact Us form.