New features

• Orama Cloud now supports returning properties. This means you can decide which properties from your documents should be sent to the browser, to reduce the data size and speed up performance at search-time.
  
  For example, if you have a list of movies and only want to return a movie title and rating:

  const data = await orama.search({
    term: 'Movie to watch with the family',
    mode: 'hybrid',
    returning: ['title', 'meta.rating']
  })

  This will reduce data transfer from the server to the browser, speeding up performance by minimizing network traffic latency.

• Added support for two new Orama Native Embedding models: orama/gte-medium and orama/gte-large. They produce vectors of respectively 768 and 1024 dimensions.

• Added support for two OpenAI embedding models: text-embedding-3-large and text-embedding-3-small. These models are now available for automatic embedding generation and usage with the Secure Proxy.

• New searchbox colors. We improved the contrast ratio of our searchbox, making it more accessible both in dark and light modes.

• Added "redeploy" function to an index: you can now redeploy an index without having to edit its configuration.

• Restyled index detail page header: added index id, implemented sticky behavior, restyled index edit description.

Fixes

• sendBeacon polyfill now works correctly in Deno.

• Fixed JSON validation error on the dashboard

• Disabled "edit mode" when deploying an index