cogibot low-level modules -- 2

We need low-level coded modules that will take and process text, sound, image, and video inputs from our existing application.
We use Azure SQL for our database.
For incoming text inputs: If an exact match is found in our existing database, stored procedures within the tables (whose structure and details we will provide to you) will be executed, and the resulting data will be returned.
If an exact match can't be found, synonym matching tables in our database will be searched word by word in an attempt to establish a match.
If a match still can't be found, the closest match will be identified (based on word similarity) and the top three closest matches will be sent to our application as a result.

Incoming sound inputs will be converted into text using voice recognition and processed in the same way as text input. Ready-to-use APIs can be utilized for voice recognition; we are open to suggestions for achieving the best results.

For incoming images and videos, text conversion will be done using OCR, and the procedure applied to text inputs will be carried out. If the images contain barcodes or QR codes, their values will be sent.

A detailed algorithm will be provided to you.