Rowbird

Run SQL on a schedule, deliver it where people read

Open Source

Analytics

Developer Tools

GitHub

Visit WebsiteSee on Product HuntGithub ⧉

Hey Product Hunt! 👋 A while ago the marketing team asked me for a report, so I wrote a Python script and scheduled it with cron. The next week they asked for another one, so I wrote another script. Then came a third, which had to go out at a different time, so that meant another cron job. Before long I was looking after a pile of scripts that nobody else could touch. One day the emails just stopped arriving. Nobody noticed at first, and when they did, I had to dig through code and logs to find out why. Maybe it was only the company I worked at, but I doubt it. That's why I built Rowbird, and why it's open source. How it works: - Write a query. The optional AI assistant can draft it for you, so it isn't only for the people who write SQL every day. - Set up the report: a schedule, a format (CSV, Excel, PDF or a table right in the message) and where it goes (email, Slack, Telegram, Discord, webhooks, S3). - Add a condition if you want, and the same report becomes an alert: "only send if there are rows", "the number went above X", "something changed since the last run". If a run fails, you know right away: failure alerts go to the channel you choose, with a fallback channel if that one fails too. Every run keeps its own history, so you're not left digging through logs. What I cared about: 🔒 It only reads your database, with read-only transactions, timeouts and row limits. It never writes. 🤖 The AI sees your schema, never your rows, and nothing runs until you approve it. 🔑 Credentials are encrypted, and it supports passkeys and SSO. No telemetry, ever. 📦 It's one small binary with SQLite, so it runs right next to the cron jobs it replaces. It isn't a BI tool. If you need dashboards, Metabase and Redash are great. Rowbird does one job: getting the right numbers to the right people on time. You can try it with Docker in about a minute: docs.rowbird.dev It's free and open source under Apache 2.0. I'd really love to hear from you: what's the first report you'd set up, and which database or destination should come next? 👇