Feature Requests

It would be nice to provide the option to do image transformations as a file is uploaded. Obviously they can be done after the fact but it would be nice if I can just do one transformation on upload (I like to just store my images as webp) then accessing the normal URL would just have the image in the right format.

Provide a download transformation to instruct browsers to download file of any type (text, binary, other), similar to the download image transformation https://xata.io/docs/sdk/image-transformations#download

Support BaaS-like real-time subscriptions. This depends on having security rules and the ability to call Xata from the browser, so https://feedback.xata.io/feature-requests/p/row-level-security is related.

It would be useful to have saved views within the table view to regularly look at and collaborate with your team. As a user you would be able to choose the columns shown, add filters (or free text search) and save the view for re-use at a later date. Taking this further, you can invite a member of your team to have access to the view and with more granular security controls, can apply RBAC to them. If we were ever to add analytics to the platform, these saved views can also double as data stores for visualizations.

It would be great if transformImage could return a format of Base64 string. This could be used with NextJS <Image /> tag's placeholder and blurDataURL.

Provide an interface for GraphQL, reflecting the Xata/Postgres schema. Include docs and examples.

Vector Search currently only returns entire documents, which can be of substantial size given that they typically store long text and embeddings. Provide a column selector to return only columns of interest.

Availability of an SDK for the creation of a distributed database via IPFS, with the possibility of configuring an IPFS-Cluster private network for data replication.

Currently, languages without spaces, such as Japanese, do not function well with the existing search engine