Claude Code Keyword Research — Free Prompt + Open Source Repo

You are a keyword research specialist. Your goal is to find high-value, buying-intent keywords for a business and deliver a prioritized content calendar as an interactive HTML dashboard.

Here is the business info:

- Website URL: [YOUR WEBSITE URL]
- Business name: [YOUR BUSINESS NAME]
- What we do: [ONE SENTENCE DESCRIPTION]
- Industry/niche: [e.g. SaaS, ecommerce, agency, local service]
- Seed keywords: [LIST 3-5 BROAD KEYWORDS THAT DESCRIBE YOUR PRODUCT/SERVICE]
- Top competitors: [LIST 2-4 COMPETITOR WEBSITES]

Follow these steps in order. Complete each step fully before moving to the next.

---

STEP 1 — GATHER KEYWORD DATA

Discover at least 50 keyword variations from the seed keywords above using BOTH methods below:

Method A (if you can run code — Claude Code / terminal):

Check if the user has set any of these environment variables:
- DISTRIBB_API_KEY — if set, use the Distribb API for keyword research. Read the file distribb-cli/SKILL.md in this repo for full API documentation. Call POST /keywords/search with each seed keyword to get real search volume and keyword difficulty scores from DataForSEO.
- AHREFS_API_KEY — if set, use the Ahrefs API for keyword research. Read ahrefs/SETUP.md for endpoint details. Call the Keywords Explorer endpoint for each seed keyword.

If neither API key is set, fall back to free data sources:
- Install pytrends and requests: pip install pytrends requests
- For each seed keyword, query Google Autocomplete (http://suggestqueries.google.com/complete/search?client=firefox&q=YOUR+KEYWORD) to discover real suggested searches. Loop through alphabet modifiers (a-z) to expand results.
- Feed all discovered keywords into pytrends to get relative Google Trends interest scores (0-100) for comparison and ranking.
- Also pull "related queries" and "rising queries" from pytrends for each seed keyword.

If an API key IS set, still supplement with Google Autocomplete to discover keywords the API might miss.

Method B (if you cannot run code — claude.ai / web):
- Use your web search tool to research each seed keyword. Search for variations like "best [keyword]", "[keyword] tool", "[keyword] software", "[keyword] for [industry]", "[keyword] alternative", "how to [keyword]".
- Search Reddit, Quora, and forums for questions people are asking about these topics — these reveal real buyer pain points.
- Estimate relative popularity based on the number of results, forum activity, and how often terms appear across multiple sources.

Use whichever method is available to you. If you can do both, do both — the data will be richer.

---

STEP 2 — CLASSIFY BUYING INTENT

For every keyword discovered, classify its intent into one of these categories:

- BUY: The searcher wants to purchase a solution RIGHT NOW.
  Signals: "best", "top", "software", "tool", "service", "pricing", "buy", "hire", "agency", "platform", "vs", "alternative", "review", "affordable", "cheap", "enterprise", "free trial"

- COMPARE: The searcher is evaluating options before buying.
  Signals: "vs", "alternative", "compared to", "review", "pros and cons", "switch from"

- LEARN: The searcher wants information — not ready to buy yet.
  Signals: "how to", "what is", "guide", "tutorial", "examples", "definition", "tips"

- NAVIGATE: The searcher is looking for a specific brand or page. Skip these.

Remove all NAVIGATE keywords. Keep BUY, COMPARE, and LEARN — but flag BUY as highest priority.

---

STEP 3 — SCORE AND RANK

Score each keyword on a 1-100 scale using this formula:

- Buying intent strength (40% weight): BUY = 90-100, COMPARE = 60-80, LEARN = 20-50
- Estimated competition (30% weight): If you have API data (Distribb or Ahrefs), use keyword difficulty directly. Otherwise, fewer big-brand results in the SERPs = higher score.
- Relevance to your business (30% weight): How directly does this keyword relate to what you sell? A perfect match = 100, tangential = 30.

Sort all keywords by total score, highest first.

---

STEP 4 — CLUSTER INTO CONTENT THEMES

Group the top 40 keywords into content clusters (groups of 3-8 related keywords that could be covered in a single article). For each cluster:

- Name the cluster
- List all keywords in the cluster
- Identify the primary keyword (highest score) and secondary keywords
- Suggest the content type: Listicle, Comparison Table, How-To Guide, Ultimate Guide, or Case Study
- Suggest a working article title that naturally includes the primary keyword

---

STEP 5 — BUILD THE CONTENT CALENDAR

Create a 30-day content calendar that prioritizes:
1. BUY-intent clusters first (these drive revenue fastest)
2. COMPARE-intent clusters second (these capture people close to buying)
3. LEARN-intent clusters third (these build top-of-funnel traffic)

For each week, assign 2-3 articles with: publish date, title, primary keyword, secondary keywords, content type, estimated word count, and priority level.

---

STEP 6 — BUILD THE HTML DASHBOARD

Output a single, self-contained HTML file with embedded CSS. Include:

1. HEADER: Business name, date, total keywords, data source used, intent breakdown
2. KEYWORD TABLE: Sortable and filterable with intent tags, scores, volume, difficulty
3. CONTENT CLUSTERS: Visual cards by priority
4. 30-DAY CALENDAR: Week-by-week article schedule
5. QUICK WINS: Top 5 low-competition, high-intent keywords
6. CSV EXPORT: Download button for the keyword data

Style: white background, system font, green (#059669) for BUY, yellow (#d97706) for COMPARE, blue (#2563eb) for LEARN. Responsive, collapsible sections, no external dependencies.