Specs
Specs turns SEO findings into work your AI coding tool can actually implement. Every improvement becomes a feature specification, with tasks, acceptance criteria, and a commit trail, instead of a bullet in a report nobody executes.
acme.comCapture work as specs#
Specs arrive two ways: you write them with New Spec, or Rampify files them for you: every site crawl that detects issues creates specs describing the fix. Filter the list by status or tag to keep planned, in-progress, and completed work separated.
Hand a spec to your AI#
The point of a spec is that you don't implement it by hand. The MCP handoff in the toolbar gives you a paste-ready prompt; your coding tool pulls the spec over MCP and works it task by task.
From the Specs toolbar; it carries the project context with it.
The AI fetches the spec with get_feature_spec and sees the same tasks and criteria you do.
As the AI marks tasks complete and links commits, the dashboard updates, with no copy-paste status meetings.
Trace specs to commits#
Tasks and acceptance criteria live on the spec; commits link back to both. Your AI can generate a conventional-commits message straight from spec context, so six months later git log still explains why a change happened.
From your AI tool
Everything above is also reachable over MCP: same data, no dashboard tab. The toolbar's MCP handoff on this page pre-fills the project context for you.