We made it because we wanted to get alerted when our Supabase tables got stale, abnormal amount of records, and anomalies in the data.
Just as an FYI, I'm intentionally writing (Soda)base for two reasons: (Soda)base's parent company is Soda Data (the data quality company), and to the untrained eye, the name Sodabase looks too similar to Supabase.
To make (Soda)base, we took Soda's ML algorithms for anomaly detection and made them super easy to set up with Supabase, so that it only takes one click. In the background, we learn the patterns in the data and the updating schedule, so when something looks off, we alert the user and fire an agent that finds the root cause of the issue.
Then the user can chat with the analysis agent in Slack to get a better understanding of it, or just hand it off to their Claude/Codex to continue the investigation or fix it.
We are also working together with Supabase to make it a one-click integration from Supabase to (Soda)base. So, if you have any feedback, questions or feature ideas, please share them here or to santi.viquez@gmail.com