seo-dataforseo

Trigger phrases

Phrases that activate this skill when typed to Claude Code:

• dataforseo
• live SERP
• keyword volume
• AI visibility check
• LLM mentions

What it does

seo-dataforseo is an extension skill in AgriciDaniel’s claude-seo repo. It connects Claude Code to the DataForSEO MCP server to provide real-time search data across 10 API modules: SERP results (Google, Bing, Yahoo, YouTube, Google Images), keyword research (volume, difficulty, intent, trends), backlink profiles, on-page analysis via Lighthouse, competitor domain intelligence, content analysis, business listings, and AI visibility — including a live ChatGPT web search scraper and LLM mention tracking across AI platforms.

The standout capability is the AI visibility module: /seo dataforseo ai-scrape <query> shows exactly which sources ChatGPT cites for a query, and /seo dataforseo ai-mentions <keyword> tracks how LLMs mention your brand across platforms. The skill includes hard cost guardrails — a cost estimation script runs before every API call and blocks calls that would exceed daily budget limits.

When to use it

Reach for it when:

• You need real search volume, keyword difficulty, or intent data rather than estimates
• You want to know which sources ChatGPT or other AI platforms are citing for your target queries
• You need live backlink data or competitor keyword rankings for a domain

When not to reach for it:

• DataForSEO extension is not installed (./extensions/dataforseo/install.sh is the setup command) — the skill will tell you this and exit cleanly
• You are on a tight budget and the cost check returns needs_approval; the skill requires explicit confirmation before any paid API call

Install

This is an extension skill requiring the DataForSEO MCP server. Install via:

./extensions/dataforseo/install.sh

Then copy the seo-dataforseo SKILL.md into .claude/skills/seo-dataforseo/.

Trigger phrases: “dataforseo”, “live SERP”, “keyword volume”, “backlink data”, “competitor data”, “AI visibility check”, “LLM mentions”, “image SERP”, “real search data”.

What a session looks like

A typical session has three phases:

1. Cost check and confirmation. Before any DataForSEO MCP call, the skill runs python scripts/dataforseo_costs.py check <endpoint>. If the result is approved it proceeds; if needs_approval it shows the cost estimate and waits for confirmation; if blocked it stops and reports the daily budget limit.
2. Data retrieval. The relevant MCP tools are called with sensible defaults (US, English, desktop unless the user specifies otherwise). Bulk endpoints are preferred over multiple single calls to minimize API cost.
3. Analysis and output. Results are formatted as tables with source labeled as “DataForSEO (live)” to distinguish from static analysis. After completing, the skill logs the actual cost and offers to generate a PDF report via /seo google report.

Receipts

Works well: The AI visibility commands (ai-scrape, ai-mentions) are genuinely novel — seeing exactly which URLs ChatGPT cites for a target query is not something any traditional SEO tool offers. For competitive intelligence, dataforseo competitors <domain> gives keyword overlap percentage and traffic estimates in seconds.

Backfires: YouTube SERP data is expensive relative to its signal value for most sites. The skill defaults to web SERP, which is the right call — but if you reflexively run YouTube searches too you can burn through daily budget quickly.

Pattern that works: Set the cost mode to threshold with a low daily limit (/seo dataforseo costs config --mode threshold --threshold 0.50) when exploring, then raise it for planned research sessions. The cost guardrails are only useful if you configure them before you need them.

Source and attribution

Originally written by AgriciDaniel. The canonical SKILL.md and supporting files live in the seo-dataforseo folder of the claude-seo repository.

License: MIT. Install, adapt, and redistribute with attribution preserved.

This page documents the skill from a practitioner’s perspective. For the formal spec and updates, defer to the source repo.