What it does
Sodabase works with Supabase projects. The reader signs in with Supabase and chooses a project. Sodabase then reads the project's tables, cron jobs, edge functions, and the logs behind them. It learns when each table usually updates and what its data normally looks like. The reader does not have to set rules for every table. When something breaks, Sodabase finds what went wrong. It points to the cron job, edge function, sync, or agent that caused the problem. It then tells the reader what to do next, and the fix arrives in Slack. One example in the product description is a signup flow that failed after a migration made a required column that a trigger did not set. In that example, Sodabase proposes updating the trigger and asks whether to go ahead.
Capabilities
- Detects stale tables, where a cron job or sync stops writing and the data is from yesterday.
- Detects schema changes, where an edge function or app expects a column that was renamed, removed, or changed.
- Detects wrong data, such as duplicated rows, spiking values, or something writing bad data.
- Learns each table's normal update timing and data patterns without per-table rules.
- Sodabase runs on the engine from Soda, a data quality company, according to the official text.
When to use it
- A signup flow broke after a Supabase migration and new users are not signing up.
- A cron job or sync stopped writing and a table holds data from yesterday.
- A table keeps updating but its numbers look wrong, with duplicate rows or spiking values.
