Browse documentation
Docs
Getting started
Connect your first repository and follow it through to a scored pull request.
- 01
Register and verify your email
Create your account at featurefactory.com, then open the verification link sent to your inbox. Verification unlocks the rest of setup and protects your organization from unauthorized access.
- 02
Choose your goal and connect GitHub
Choose the outcome you want to improve, then connect the GitHub App from the setup screen. FeatureFactory uses read-only access for measurement, and you decide which repositories are included.
- 03
Pick your repositories
After GitHub is connected, select at least one repository from the repository picker. Activating a repository automatically starts its first history backfill. You can change the selection later in Settings.
- 04
Follow the history backfill
FeatureFactory imports pull requests and commits in the background. The setup screen advances as history arrives, and the repository settings page shows progress and recovery actions. The available history window depends on your plan and repository data.
- 05
Review your first scores
As soon as scored pull requests are available, open Measure to review velocity, quality, and AI attribution. Each pull request detail page explains its score, effort rationale, quality evidence, and attribution signals.
- 06
Try Build or finish setup
Organizations with the Build pillar can start an agent-driven run from an approved plan or workflow template. You can also skip this step, open the dashboard, and invite teammates from Settings > Members when you are ready.
Common questions
How much history is imported?
The available history window depends on your plan, repository history, and provider data. Repository settings show the active measurement window and sync status.
Does FeatureFactory write to my repositories?
No. The GitHub App is read-only. FeatureFactory never pushes commits, creates branches, or modifies pull requests.
When are new PRs scored?
GitHub notifies FeatureFactory after a pull request closes and scoring begins automatically. Queue load, repository size, and configured analysis can affect processing time.
How do I instrument my AI agents so FeatureFactory detects them?
Add Co-Authored-By: <agent-name> <noreply@agent> trailers to commits your agents create. See AI attribution for the full guide.
Was this article helpful?
Your response enters the same customer insight queue used by product feedback.