I'm building Voxoria (https://voxoria.ai), it tracks whether B2B brands get mentioned when people ask ChatGPT/Perplexity/Gemini instead of Googling.
Ask the same engine the same question twice and you get different answers, different citations, sometimes a different opinion of your brand, so figuring out how best to present this has been a fun product problem to solve.
It also tries not to be yet another dashboard: instead of just analytics, an agent turns the findings into a ranked list of "ship this fix" todo items.
It's an incident management platform, similar to Pagerduty, Rootly or FireHydrant.
It's the first side project I've open sourced, and I've been hacking on it weekends and nights. Hoping to get a few companies to start using it to get some early feedback.
Using LLMs to classify noisy alerts is a really clever approach to tackling alert fatigue! Are you fine tuning your own model to differentiate between actionable and noisy alerts?
I'm also working on an open source incident management platform called Incidental (https://github.com/incidentalhq/incidental), slightly orthogonal to what you're doing, and it's great to see others addressing these on-call challenges.
Our tech stacks are quite similar too - I'm also using Python 3, FastAPI!
We used to be customers of Mailgun, but last month moved over to SES. Not that there's anything wrong with Mailgun, SES's pricing is far cheaper (60K free emails if you're sending from an AWS instance), and deliverability is on par.
But SES is super basic. It's just an API, no fancy analytics.
So of course, I built a small web app that "wraps around" SES, and gives you some insight into what you're sending, some pretty graphs, and other stuff.