Skip to content

For product and growth

You want to change how the product behaves without waiting for an engineering release: flip a feature on for beta users, try a new prompt on 10% of traffic, or roll something back the moment it looks wrong. In Logfire, the tool for this is feature flags (Logfire calls them managed variables): configuration you define once in code but control from the Logfire UI, without redeploying.

Each link says why it’s here.

Your path

  1. Feature flags: define a value once in your code with a sensible default, then change it live from the Logfire UI. A flag can be a simple on/off switch: turn a feature on or off, or target it to a specific group (beta opt-ins, internal users, a single customer) using your existing trace attributes as the targeting rules; or it can hold richer configuration such as a prompt.

  2. Feature flags for web and mobile apps: serve those same flags to browser, mobile, and edge clients through the OpenFeature Remote Evaluation Protocol (OFREP), an open standard for reading feature flags from any client, so any OpenFeature-compatible app can read them without the Python SDK.

  3. Tie changes back to usage: every managed-variable resolution is recorded in your traces. That means you can correlate a flag or prompt change directly with what happened next (the basis for A/B tests and online experiments) using the same data your app already sends. See Explore to query that data in SQL and Dashboards to watch it over time.

Try the full journey

  • Roll out a prompt safely: change a live prompt for some traffic, watch the effect in your data, and roll it back if it gets worse.