# DataForSEO Integration Guide

## Overview

DataForSEO provides real keyword data, competitor intelligence, and AI search optimization metrics. This replaces guesswork with data-driven decisions.

**MCP Access:** DataForSEO tools are available through MCP. Use them directly in your research workflow.

## Budget Protocol

- **Per session limit:** $0.50 (unless user explicitly approves more)
- **Monthly budget:** ~$10
- **Always report:** Show user what API calls you're making and estimated cost

## Core Principle

Start with seeds → expand with related → filter by opportunity → verify with SERP.

Don't chase high-volume competitive keywords. Find gaps where we can win.

---

## For @spy: Competitive Intelligence

### Competitor Keywords
```
Tool: dataforseo_labs_google_ranked_keywords
Use: See what keywords competitors rank for
Target: fal.ai, replicate.com, runware.ai, cloudinary.com
```

### Backlink Analysis
```
Tool: backlinks_summary, backlinks_referring_domains
Use: Where competitors get links, potential outreach targets
```

### Domain Intersection
```
Tool: dataforseo_labs_google_domain_intersection
Use: Find keywords multiple competitors rank for (validated demand)
```

### LLM Mentions (GEO)
```
Tool: ai_optimization_llm_mentions_search
Use: Check if Banatie or competitors mentioned in AI responses
Platform: chat_gpt, google (AI Overview)
```

---

## For @strategist: Keyword Research

### Search Volume
```
Tool: keywords_data_google_ads_search_volume
Use: Get real monthly search volume for keyword list
Input: Up to 1000 keywords per request
```

### Keyword Difficulty
```
Tool: dataforseo_labs_bulk_keyword_difficulty
Use: Score 0-100, lower = easier to rank
Filter: KD < 50 for realistic targets
```

### Related Keywords
```
Tool: dataforseo_labs_google_related_keywords
Use: Expand seed keywords, find long-tail opportunities
Depth: 1-4 (start with 1, go deeper if needed)
```

### Search Intent
```
Tool: dataforseo_labs_search_intent
Use: Classify keywords as informational/navigational/commercial/transactional
Match: Content type should match intent
```

### AI Search Volume (GEO Priority)
```
Tool: ai_optimization_keyword_data_search_volume
Use: Keywords popular in AI search (ChatGPT, Perplexity)
Why: Early indicator of emerging queries
```

### Research Workflow

1. **Start with seeds** (3-5 per topic)
2. **Get search volume** for seeds
3. **Expand** top 3 by volume with related keywords
4. **Filter:** Volume > 50, KD < 50
5. **Check intent** for finalists
6. **SERP analysis** for top candidates

---

## For @seo: Optimization & Verification

### SERP Analysis
```
Tool: serp_organic_live_advanced
Use: See current top 10 results, SERP features present
Check: Featured snippets, PAA, video results
```

### On-Page Analysis
```
Tool: on_page_instant_pages
Use: Technical SEO check of specific URL
After: Publishing, verify optimization
```

### LLM Responses (GEO)
```
Tool: ai_optimization_llm_response
Use: See how AI models answer our target queries
Why: Optimize content for AI citations
```

---

## Key Learnings

**Problem-aware keywords often have zero volume.**
People search for solutions, not problems. "placeholder images slow" = 0 volume. "generate images api" = real volume.

**Related keywords > seed keywords.**
Your initial guesses are rarely the best targets. Let data guide expansion.

**Brand keywords are useless.**
"cloudinary pricing" means they already chose Cloudinary. Target problem/solution queries.

**Low KD + decent volume = opportunity.**
Don't chase "ai image generation" (KD 80+). Find "generate images for nextjs" (KD 30, volume 200).

---

## Output Format

When reporting keyword research:

```markdown
## Keyword Research: [Topic]

### Seeds Analyzed
| Keyword | Volume | KD | Intent |
|---------|--------|----|----|
| ... | ... | ... | ... |

### Top Opportunities
| Keyword | Volume | KD | Rationale |
|---------|--------|----|----|
| ... | ... | ... | Why this is a good target |

### Recommendations
[What content to create based on this data]

### API Calls Made
[List of tools used, estimated cost]
```