Documentation
Concepts & data model
The four-object model and the six-question loop that Slope is built around.
How Slope thinks
Slope is built around four objects and one loop. Understand these and you understand the product.
Organization
A multi-tenant boundary. Every piece of data — project, integration, opportunity, task — is scoped to an organization_id and never crosses. Members + roles live here. Billing happens here. When you sign up, Slope auto-creates your personal organization.
Project
A single website (domain) you want to grow. A project has a domain, competitor list, and connected integrations (GSC, GA4, GitHub). All opportunities, pages, crawls, and tasks belong to a project.
Opportunity
A detected growth fix. Slope's seven rule families produce these. Each has a title, evidence (the data that triggered it), recommended actions, and a priority_rank computed from impact × confidence ÷ ease.
Execution task
A GitHub issue Slope created for an approved opportunity. Linked back to the opportunity by id. When the issue closes in GitHub, Slope marks it completed. 14 days later, the impact worker reads GSC and writes the actual clicks/impressions/position delta to the task.
The loop
- Connect — GSC, GA4, GitHub, optionally LLM keys.
- Sync — daily orchestrator pulls fresh metrics + crawls new pages.
- Detect — rule engine evaluates against latest data, surfaces opportunities.
- Triage — you review, approve, ignore, or push to GitHub.
- Execute — your team (or AI coding agent) ships the fix.
- Measure — Slope reads GSC 14 days later, writes the result back.
Every feature in Slope answers one of those six steps. If a feature doesn't, we don't build it.