What is it?
Redash is an open source business intelligence tool with a Python backend and a JavaScript and TypeScript frontend. Every chart starts as a query: you write it in SQL, or in the source's own query language for MongoDB or Elasticsearch, run it, then add a visualization to the result and pin it to a dashboard. The README spells out the division of labor: people who write SQL explore and visualize, and the rest of the company reads what they built.
That approach sets Redash apart from Metabase, whose visual query builder is aimed at people who don't write SQL, and from Apache Superset, where charts sit on datasets and metrics defined once for the whole team.
The comparison of open source BI tools weighs the state of Redash, now run by its community, against the other tools.
Why is it interesting?
- Query editor: schema browser, autocomplete, and parameters in double curly braces (
{{ }}) so the same query can run against another date or value without being edited. A dropdown parameter can pull its options from another query. - Data sources: the built-in list in the README has 69 entries. PostgreSQL, MySQL, BigQuery, Snowflake, ClickHouse and Trino are there, along with Prometheus, Google Sheets, CSV or Excel files and JSON APIs.
- Cross-source joins: the "Query Results" source runs SQL over the results of other queries in an in-memory SQLite database, so one query can join a PostgreSQL table with a Google Sheets tab.
- Scheduled refreshes and alerts: a query can rerun on a fixed interval, and an alert fires when its result meets a condition. The repository ships twelve alert destinations, including email, Slack, Microsoft Teams, PagerDuty and webhooks.
- Sign-in options: the code handles Google OAuth, SAML, LDAP, JWT tokens and identities passed by a reverse proxy.
Good to know
A community-run project. Databricks acquired the company behind Redash in June 2020. In April 2023, Arik Fraimovich, the project's author, rebooted it as a community-led effort with seven volunteer maintainers, keeping the BDFL role himself. In July 2023 he added that Databricks, which could no longer maintain the project, keeps an internal fork built into its platform and provides no support for developing the open source version.
Three years without a release, then two a year. Nothing shipped between 10.1.0 in November 2021 and 25.1.0 in January 2025. Version numbers now follow the year and month: 25.8.0, 26.3.0, then 26.9.0 on September 24, 2026. That release fixes thirteen published security advisories, six of them rated high, and turns off custom JavaScript visualizations by default. The latest commit is from October 3, 2026.
What you have to run. PostgreSQL for metadata, Redis, a web server, a scheduler and workers that execute the queries. The reference Compose file, in the getredash/setup repository, starts eight containers, including three workers and an nginx. The docs recommend at least 4 GB of RAM for a basic deployment, and every install method they describe goes through Docker. "Query Results" joins happen in memory and can fail on large result sets.
One server to start with. The setup script's FAQ calls it fine for production on small deployments. Beyond that, it recommends moving PostgreSQL, and probably Redis, to their own servers and running Redash on at least two servers.
Restricted viewers get typed parameters only. A text parameter requires full access to the data source and cannot be used in shared dashboards, because of the SQL injection risk. View-only users are limited to typed parameters: number, date, dropdown list.
A permissive license. Redash is BSD-2-Clause, where Metabase is AGPL-3.0 and Superset is Apache-2.0. The redash.io site lists no pricing and no hosted plan.