This thread will continue to serve as a resource hub for follow-ups.
Have questions or insights? Drop them in the thread!
(Original Post)
Ever wonder how to actually use different AI models together in your apps? Tomorrow, we’re hosting a live, hands-on session, and we’d love for you to join us!
What we’ll cover:
Experiment with LLMs from OpenAI, Anthropic, and Google Build multi-model applications that leverage different strengths Implement enterprise-grade security with Amazon Bedrock or internal proxies Access 200+ models via OpenRouter with intelligent request routing
AI models are evolving fast, and so are the ways we build with them. Whether you’re already experimenting or just getting started, this is a great chance to swap insights and walk away with new ideas to try in your own projects.
Hey y'all! Excited to chat with everybody. Feel free to drop questions here in advance (or if you can't make it) and I'll do my best to either include them in the presentation or answer them here!
Thank you so much for your time, Keenan, really enjoyed learning more about the background context and more basic information about LLMs/AI that I definitely didn't know.
Question:
Does AI work well to analyze data in a Retool postgres database?
Example 1: how many visits (a table) did Technician A make last week?
Example 2: how much revenue did we generate from Client A last month for Property B?
Example 3: What percentage of quotes did Teammate A close last quarter?
Yes, LLMs to a pretty good job at going from natural language to database query.
So in this example, you would ask the question, the LLM would convert "how many visits (a table) did Technician A make last week?" into "SELECT * FROM visits WHERE technician = "techA"" and then you could use a standard RetoolDB query to pull that data out.
Fantastic session!@WidleStudioLLP is always exploring innovative ways to integrate AI into internal tools, and the insights on combining multiple LLMs and leveraging OpenRouter for model routing were truly valuable. Excited to implement these strategies to enhance AI-driven Retool apps. Huge thanks to the Retool team for the great discussion!
What a great session! Thank you to everyone who joined, asked thoughtful questions, and shared feedback—it was awesome to see so much engagement. I'm glad many of you found it helpful!
I've updated the original post with all of the information you'll need to dive back into the material. Happy learning!
Hey all! In addition the recording posted above, here are some resources from our discussion yesterday.
Chat with PDF app
This was the first example we showed where you can upload a PDF and get a summary or have a chat and ask questions about it. You'll want to make sure the AI query is hooked up to the AI in your own Retool account, but this will give you a place to start.
This app lets you compare two LLMs side by side, with only having to enter one prompt. Again, make sure the AI queries are hooked up and using the models you want.
You can find the inspiration for this at: https://lmarena.ai/, but our custom Retool version is here. Once you get the app imported, in addition to hooking it up to your own AI resource, you'll need to set up two RetoolDB tables: chatbot_arena_matches and chatbot_arena_models. Here are two CSVs of sample table data so you can get started.
UPDATE
Model Mix and Match: How to Build AI Apps with (Almost) Any LLM
Resources from the session – example apps, setup instructions, and key takeaways.
Have questions or insights? Drop them in the thread!
What we’ll cover:
Experiment with LLMs from OpenAI, Anthropic, and Google
Tomorrow, March 4 at 10 AM PST
in the comments if you're joining!
Fantastic session! 
Thank you to everyone who joined, asked thoughtful questions, and shared feedback—it was awesome to see so much engagement. I'm glad many of you found it helpful!