Workbench technology strategy

1. Basic client – server approach

   The workbench uses a Javascript enabled web browser as the client for the User Interface.

   The user interface is developed around the Node.js and React tools

   The backed server is written in python as a FastAPI multi-threaded engine that provides AI access services, statistical analysis functions, and visualisation services for the client layer

   2. Workbench UI features

      The workbench has some fundamental features

      • Production and Demo modes

      • Multi lingual (currently, English, French, German, Spanish and Thai)

      • User profile and access control

      • Self Registration of users

      • Self Management of Database connections – multiple per user

      • Self Management of AI connections – multiple per user

      • Self Management of Analysis recipes – multiple per user

      • Admin management of system limits and users

      • Admin control of allowed AI providers and models

      3. Workbench platform integrations

         The workbench directly supports connections to a growing list of other platforms and local servers.

         1. Database providers

            • PostgreSQL server – localhost or remote

            • MySQL (or MariaDB) server – localhost or remote

            1. AI providers

               • Ollama locally

               • a lot of Remote OpenAI accessible platforms

               4. The basic UI layout (including Admin functions) looks like this

                  O ne easy AI component is the usual sort of Chat interface ..

                  
                  
                  

                  This is a little different, because it is a unified chat with any Provider/Model that you have defined.

                  The available AI Providers are controlled by the admin, and some Models may be disabled by an admin

                  4. ;padding-top:0"> ← Back to Table of Contents