Carry flag cleanup into code
Connect
SPIRITT
StatsigStatsig
Carry experiments through to decisions
SPIRITT carries Statsig experiments from setup to evidence-backed decisions. Coordinate rollout reviews, inspect real metric results and turn retired flags into code cleanup work that reaches the engineering team.
What you can do with Statsig.
Rollouts with a record
Create flags and experiments, route proposed changes through eligible reviewers, and distinguish approval from the live commit.
Decisions grounded in results
Combine overall experiment outcomes with metric breakdowns and version history before changing exposure.
Retire the code, too
Start cleanup pull requests for stale gates and experiments and carry the review handoff into your engineering workflow.
Works with your other apps.
Get started in three steps.
01
Open your workspace
Start a SPIRITT workspace for experiment operations with Statsig.
02
Connect Statsig
Connect your Statsig account and choose the account access your mission needs. Keep other connected apps scoped to the same work.
03
Delegate an outcome
Describe the result, the Statsig records to use, and the rules for changes or delivery. SPIRITT builds the workflow and follows the result through.
Hand over a mission.
Choose a mission. Make it yours.
Put Statsig to work.
Carry experiments through to decisions
Questions
Statsig, with SPIRITT.
01Does approval make a Statsig change live?+
No. Reviews separate proposal, approval and commit. SPIRITT tracks that lifecycle and verifies the gate, experiment or config after the authorized commit.
02Can it decide who actually passes a gate?+
It can inspect effective configuration. Enabled state and status take precedence over rules: a launched or disabled gate can return a blanket value rather than evaluate targeting rules.
03Can it compare experiments beyond a headline result?+
Yes. Overall results, individual metric breakdowns, metric definitions and version history support a decision brief with the relevant cohorts and configuration changes.
04Can it remove stale experiment code?+
It can start Statsig cleanup pull requests for gates and experiments, then coordinate review in GitHub. A created pull request is not a merged cleanup.
05Can it manage AI prompts?+
Yes. SPIRITT can create prompts, add versions, update metadata and start offline evaluations. Evaluation submission is tracked separately from a completed result.
06Which access is needed for reviews?+
Review actions need suitable user-owned credentials and eligible reviewers. SPIRITT follows the project's review rules rather than assuming authors can approve their own changes.