Stop spending money & time on replicas/caching/microservices
when the fix was a one-liner. Happens *all the time*.
dbpill intercepts, finds & fixes sneaky slow queries
Buy standalone executable
$100 Free
- → One-time fee, updates included
- → OS: Mac, Linux & Windows (x64 & arm64)
- → DB: PostgreSQL only
Usage
$
./dbpill postgresql://user@host:5432/db
PostgreSQL proxy running on port :5433
→ Connect to dbpill SQL proxy at postgres://user@localhost:5433/db to intercept queries.
→ Go to dbpill web UI at http://localhost:3000 to manage the results.
█
No code changes required Just point your backend's SQL connection to dbpill.
One-click AI analysis Use your own LLM API keys with the latest models.
No guesswork, exact performance metrics See the performance of your queries before & after indexing.
One-click apply, one-click revert Instantly test suggested index improvements.
Quick Tips
- Works best with a populated database; using a clone of your production database is recommended.
- Most queries should take under a millisecond when indexed correctly, even on tables with millions or even billions of rows.
- Even if you have small tables, you can apply suggested indexes early to avoid future performance issues.
Privacy
dbpill is a local application. All intercepted queries are processed on your computer, and saved to a sqlite database in your filesystem (next to the executable).
No data is sent to our servers. In fact, there are no servers.
LLM usage is purely optional. If you choose to use LLMs to assist with optimization, you can pick your own providers and API keys.
The LLM prompts are public and can be previewed. You can view them in the web UI before they are sent.
If you have additional security & privacy requirements, please contact sales@dbpill.com for enterprise support.
FAQ
How is this different from slow query logs? In addition to slow queries, dbpill also intercepts query params which allows it to re-run the queries with EXPLAIN ANALYZE to get full query plans. This is not possible with slow query logs.
What happens when an INSERT or UPDATE query is re-run? EXPLAIN ANALYZE on updates or inserts do not modify the table, they only return the query plan. It's safe to re-run them as many times as you want.
What if suggested indexes don't improve performance? You may need to modify the query, which is outside of dbpill's current scope.