Releases

Version 1.2.x was all about Kafka implementation, and It's here now! Neura Query is now using Kafka to support robust architecture, allowing scaling across different servers very easily. FastAPI and ServerManger are now separate services. Accepting requests and handling them are now different processes. Every user request to Neura Query is now delivered to FastAPI, registered on a PostgreSQL DB, and being handled by the ServerManager. Kafka is responsible to push messages to the ServerManager.

Logger fixed, now it supports Hebrew too.

Fixed AI prompts for better responses.

DB URI connection is now saved on NeuraQuery servers. On next login, NeuraQuery will remember the last DB URI the user was working with.

AI prompts improved. Now the AI is more consistent with its responses.

Added logger. Now Neura Query will log queries and commands on the API, enabling convenient way to debug errors, like DB commands failing when they shouldn't.

No need to supply provider name anymore. Neura Query will figure out by itself what is the SQL engine type, and use to create queries accordingly.

New design for connection page. Now you can use a demo DB to work with the AI.

New design for query window. With a new looking design, now it is much more convenient to query the AI and read the results.

Enhanced AI prompts are on the go. Now Neura Query can support a wide variety of human queries. It can actually support general questions about your DB and response in much more elaborate manner.

New update! Now Neura Query will refresh its metadata cache whenever it is changed.

Client now receives and updates GUI with DB commands and their results automatically.

Metadata and provider information moved from cookies on client side to api server. This move makes fetching this information much faster by making it accessible from the api server itself. It also reduces the payload size sent in every request. And most importantly, cookies has limitations of 4096 Bytes size, meaning metadata cannot be stored if it is to big, moving to api server solves this problem.

Neura Query is released for the first time :).