banatie-content/project-knowledge/research-tools-guide.md

242 lines
6.2 KiB
Markdown

# Research Tools Guide
## Overview
Three research tools available through MCP:
| Tool | Best For | Cost |
|------|----------|------|
| **DataForSEO** | Structured SEO data (volumes, KD, SERP features) | Paid (~$0.50/session) |
| **Brave Search** | Fast web search (news, Reddit, competitors) | Free |
| **Perplexity** | AI synthesis ("what's known about X") | Free |
**Strategy:** Use free tools liberally for discovery. Use DataForSEO strategically for validation.
---
## Tool Distribution by Agent
| Agent | DataForSEO | Brave Search | Perplexity |
|-------|------------|--------------|------------|
| @spy | ✓ keywords, backlinks, LLM mentions | ✓ news, Reddit, HN | ✓ deep research |
| @strategist | ✓ volumes, difficulty, intent | — | ✓ content landscape |
| @seo | ✓ SERP, on-page, LLM responses | ✓ what ranks now | — |
| @webmaster | — | ✓ competitor pages | ✓ messaging research |
---
## Brave Search
### When to Use
- Breaking news about competitors
- Community discussions (Reddit, HN, Twitter)
- What's currently ranking for a keyword
- Competitor content examples
### Query Patterns
```
"runware ai news" → competitor updates
"site:reddit.com ai image api" → community pain points
"site:dev.to placeholder images" → existing content
"replicate.com pricing" → competitor pages
```
### Example Workflow (@spy)
```
1. brave_search: "runware ai" → recent news
2. brave_search: "site:reddit.com mcp image generation" → community sentiment
3. Synthesize findings into research/*.md
```
---
## Perplexity
### When to Use
- Understanding what's already written about a topic
- Getting synthesized overview of a domain
- Deep research questions
- Competitive positioning analysis
### Query Patterns
```
"What tutorials exist about Next.js image optimization" → content landscape
"How do AI image APIs position themselves to developers" → messaging analysis
"What are developers saying about MCP servers" → sentiment synthesis
"Comparison of placeholder image services" → competitive intel
```
### Example Workflow (@strategist)
```
1. perplexity: "What content exists about AI placeholder images" → landscape
2. DataForSEO: keyword research for gaps → validate demand
3. Decision: write or skip
```
### Example Workflow (@webmaster)
```
1. brave_search: "replicate.com pricing page" → see competitor pages
2. perplexity: "How do AI APIs explain pricing to developers" → messaging patterns
3. Create pages/*.md with informed positioning
```
---
## DataForSEO
### Budget Protocol
- **Per session limit:** $0.50 (unless user explicitly approves more)
- **Monthly budget:** ~$10
- **Always report:** Show 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 research:
```markdown
## Research: [Topic]
### Tools Used
- Brave Search: [queries]
- Perplexity: [queries]
- DataForSEO: [tools, estimated cost]
### Findings
[What you discovered]
### Keywords (if applicable)
| Keyword | Volume | KD | Intent |
|---------|--------|----|----|
| ... | ... | ... | ... |
### Recommendations
[What to do next]
```