Product docs
How SerpGoblin Works
SerpGoblin turns search performance data into a ranked work queue, then keeps the brief, handoff, and outcome measurement attached to each SEO issue.
Operating Model
SerpGoblin is built around one loop: connect trusted search data, surface the next useful SEO fix, prepare a brief an operator or coding agent can act on, and measure the result after fresh data arrives.
Connect Signals
Google Search Console provides queries, pages, clicks, impressions, CTR, and position. Analytics can add post-click context.
Rank Work
Analysis turns raw signals into issue cards sorted by likely lift, confidence, effort, and current lane.
Track Outcomes
Fixed issues move into Tracking so the next data window can compare expected movement against observed search performance.
Core Concepts
- Project
- A site or property you want SerpGoblin to analyze and track.
- Property Selection
- The Search Console property connected to a project for analysis.
- Issue Card
- A ranked SEO opportunity with evidence, rationale, and workflow state.
- Brief
- A ship-ready implementation summary generated from the issue evidence.
- Tracking
- The state for fixed issues waiting for enough fresh data to measure impact.
Board Lanes
Issues that are ranked and ready to review.
Issues currently being expanded into implementation briefs.
Briefs that are ready for an operator or agent to inspect.
Resolved work that should be measured after the next data refresh.
Where AI Agents Fit
The MCP server exposes the same board state to coding agents that users see in the app. An agent can list projects, inspect the board, read one SEO issue, prepare a brief, and mark the work resolved once a fix ships.
https://serpgoblin.com/api/mcp
The endpoint uses Streamable HTTP and OAuth bearer authentication. The client discovers the authorization server from the protected resource metadata exposed by this application.