Use AI to chat with a SQL Database 5.0 - AWS Bedrock AI, GitHub AI Models, Oracle Databases

DbChatPro is a free app that let's you query and chat with your relational database using AI and natural language. The new 5.0 version introduces support for AWS Bedrock's AI service and models, as well as GitHub AI models, and like previous versions still of course supports OpenAI, Azure OpenAI, and local AI models. On the database side, the 5.0 version adds support for Oracle relational databases, and continues to support SQL Server, PostreSQL, MySQL. There's also now an option to export your data table results as CSVs.

This video also covers some exciting changes to the DBChatPro project itself - including improved readme setup and configuration instructions, a separate video entirely dedicated to app setup and configuration, and a unified GitHub repository and release strategy going forward.

Sample app repo:

View the full playlist here:

Related: Use AI to chat with a SQL database v4.0

00:00 Intro
01:52 New features & app demo
07:47 Set up the app
12:14 Code concepts
17:48 Project news & updates

My other AI and coding videos:

Build a fully local AI RAG solution with .NET

The .NET AI Ecosystem as simple as possible

Build .NET AI Apps using Microsoft.Extensions.AI - the future of .NET AI!

.NET 9.0 Blazor Features in 9 Minutes

Setup Stable Diffusion locally:

AI Upscale any video on your local machine

[ventoshe]

Incredible project, excited to see the new features and improvements as it evolves! 🙏🏆

[Bayzon]

Another great release, thank you! I’m implementing something similar at work and this series has been eye opening

[ironman-wq]

hi everyone who every feeling this was amazing yes it is.
But there is one issue it don't work for the big data which has the large number of tables and columns and pk's and fk's.
before doing analysis you need to optimize your database normalize the db then only it works well.
This is approach is good only for the data which is for small like as he provided in the sample above and which the llm even understand easily.
The llm also hallucination so the data may sometimes provide error sql dont work all the time.

[queenstownswords]

This rocks. Take my like Sir... and take the star on the repo! Good day to you Sir.

[dheerajgupta]

Awesome content.
Just one request if you can create video to set up this app with AWS bedrock

[prashwilliam]

First of all, great video again ! How would you take care of scenarios where the schema is not well defined ? For eg For an Employee table we created the table name with Emp instead of Employee? Also, what do we do if there are 100's of such tables ? What prompt can we provide ?

[DougNelson]

Great video, I am looking forward to trying it out. One question, are there any guardrails set up? What happens if I ask it to DROP ALL TABLES?

[vmrosario]

Great job on the application! I’m curious about how it infers relationships between tables—does it rely solely on the schema, or does the AI model handle it automatically? Also, would it be possible to incorporate stored procedures, views, and functions so the AI can assist in creating new ones or understanding table relationships more effectively in SQL server. Thanks!

[jkcootori]

Do we need to setup other things if we use Github token ?, and can we only allow only read actions for the ai ?

[MuditMehrotra-wp]

Can we add Admin related tasks to the agent like creating database, drop database etc.

[vinayrk]

Could you update the DB connection string document to reflect the connection string for Azure SQL and Azure Mysql. Its become a pain to make this work.

[AddictedtoGaming]

HELLO BRO I NEED THIS APP HOW CAN I GET IT.

[nicolassuarez]

Ok, does the job, but if you have to create complex queries LLMs have no clue how to do it, even knowing the schema...

[UzairAli]

Looks like you forgot to make this repo Public