Allows pushing generated SEO plan items into a Notion board for task tracking and management.
Point it at your site, get a whole-site AEO/GEO audit plus a ready-to-run plan your agent can execute. One magic button for AI-SEO - runs as a Claude skill, a Claude plugin, and an MCP server, so it works in any MCP host (Cursor, Cline, Windsurf, Claude Desktop).
It produces an actionable plan, not direct edits: a portable plan.json plus
a markdown checklist, where every item is self-contained and agent-executable.
You then run your agent against the plan to execute it.
Under the hood it is a thin orchestration layer over two engines:
@automatelab/ai-seo-mcp
(audit + score + rewrite) and
@automatelab/citation-intelligence
(what AI engines cite). No API keys are needed for the core audit→plan flow.
Run it → it crawls your site, audits every page, and generates
plan.json + a markdown checklist, prioritized by estimated score lift.
Run your agent against the generated plan to execute it (rewrites, schema, llms.txt, robots) - then verify the score actually lifted.
Related MCP server: automatelab-ai-seo
// e.g. claude_desktop_config.json / Cursor / Cline
{
"mcpServers": {
"ai-seo-magic-button": {
"command": "npx",
"args": ["-y", "ai-seo-magic-button", "mcp"]
}
}
}Exposes one tool: generate_seo_plan { domain, pages? } → returns the full
plan.json (as structured content) plus the markdown checklist.
/plugin install AutomateLab-tech/ai-seo-magic-buttonCopy skill/ into your .claude/skills/ (or install via the plugin above). The
skill is the conversational magic-button UX - see skill/SKILL.md.
npm install && npm run build
# audit -> plan.json + plan.md
node dist/cli.js audit example.com --pages 10
# execute the plan's tool-driven items (writes to ./magic-button-out)
node dist/cli.js apply plan.json
# re-audit and diff the score (proof the plan worked)
node dist/cli.js verify plan.json
# optional: push items into an agency-os Notion board (else local checklist)
node dist/cli.js sink plan.jsonThe child engines resolve via npx by default. To use local builds:
export AISEO_MCP_PATH=/abs/path/ai-seo-mcp/dist/index.js
export CITATION_MCP_PATH=/abs/path/citation-intelligence-mcp/dist/index.js$ node dist/cli.js audit example.com --pages 2
https://example.com
pages audited: 1 · avg score: 48 (D)
fixes: 8 · est. lift: +78 points
1. [critical] +12 No JSON-LD structured data found on this page.
2. [critical] +12 No FAQ structure found (no FAQPage schema or H3 question headings).
3. [critical] +12 No sitemap found at standard locations.
6. [high] +10 Rewrite for AEO: this page
7. [medium] +4 Low authority signals - missing Organization/author schema.
8. [medium] +4 No canonical link element found.A full sample plan.json is in examples/plan.example.json.
Each item carries a self-contained action (the tool + baked params) and an
acceptance check - see docs/plan-format.md.
One shared core lib, four thin surfaces. See
docs/scope.md for positioning and
docs/plan-format.md for the schema + core-lib boundary.
src/core/ audit + plan + apply + verify + sink (the only place with logic)
src/cli.ts CLI surface
src/mcp-server.ts MCP server surface (generate_seo_plan)
skill/ Claude skill surface
plugin/ Claude plugin surfaceMIT
Crawl a site, run a whole-site AEO/GEO audit (via the ai-seo + citation-intelligence engines), and return a portable, agent-executable plan.json plus a markdown checklist. This produces an actionable PLAN, not direct edits.
| Name | Required | Description | Default |
|---|---|---|---|
| pages | No | Max pages to audit (default 10) | |
| domain | Yes | Site to audit — hostname or origin, e.g. example.com | |
| include_info | No | Include purely-informational findings (non-actionable notes). Off by default. |
| Name | Required | Description |
|---|---|---|
| items | Yes | |
| domain | Yes | |
| generated_at | Yes | |
| pages_audited | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnly, idempotent, openWorld, and non-destructive; the description corroborates this by clarifying that it only crawls and produces a plan, not edits. It also names the two output artifacts and the underlying engines (ai-seo + citation-intelligence), adding context beyond the annotations. It stops short of disclosing crawl duration or rate/scale considerations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two tight sentences with no filler; the action and the returned artifacts are front-loaded, and the plan-vs-edits caveat is placed where it will be read. Every clause earns its place.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With a full schema, complete annotations, and an output schema, the description's burden is low, and it still names both output artifacts. It is essentially complete for correct invocation; only runtime/scale expectations for auditing up to 50 pages are left unaddressed.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100% with only three simple parameters, so the schema already explains domain, pages (with default and max), and include_info. The description adds no parameter-level meaning beyond that, which is the correct baseline when the schema carries the load.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
States a concrete verb chain (crawl, audit, return) plus the exact deliverables: a portable agent-executable plan.json and a markdown checklist. The clarifying clause 'This produces an actionable PLAN, not direct edits' removes any ambiguity about the operation's nature, which matters since the name alone could imply mutation.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It clearly frames the context of use ('whole-site AEO/GEO audit') and explicitly delimits the outcome as a plan rather than applied edits, which tells an agent when this tool is the right choice. There are no sibling tools to differentiate against, so the absence of named alternatives is not a gap, but no explicit when-not-to-use condition is given.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.0generate_seo_planScored across 1 tool
There is only one tool, so no selection ambiguity is possible. Its purpose (crawl + full-site AEO/GEO audit + generate plan.json and checklist) is stated clearly and distinctly.
The single name `generate_seo_plan` follows a clean verb_noun snake_case convention. With one tool there is no risk of mixed conventions.
One tool for an entire 'SEO Magic Button' server is thin, sitting at the low end of the 3-15 range. It is offset by the tool being substantive (crawl + multi-engine audit + plan generation in one call), but there is still only a single entry point.
The stated purpose of producing an actionable plan is covered, but the surface is a one-shot generator with no follow-up operations such as retrieving/list existing plans, checking audit status, or applying/executing the generated plan. The description explicitly disclaims direct edits, so agents may hit a dead end after planning.
Open-source AI SEO over MCP: audits, ranks, keywords, backlinks + AI visibility (GEO).
SEO MCP server: crawl your site, find AI-visibility gaps, and ship the fix from your coding agent.
Run SEO + AI-visibility (GEO) audits from Claude, Cursor & other AI clients.
Free technical-SEO audit MCP: crawl a site, run checks, return an LLM-ready shareable report.
