How to Do Keyword Research From Your Terminal (No Browser Required)

Most keyword research starts with opening Ahrefs, Semrush, or Google Keyword Planner in a browser. You type a seed keyword, wait for results, export a CSV, then manually process the data.

If you’re a developer, that workflow feels broken. You live in the terminal. Your content lives in markdown files. Your deployment is a git push. Why should keyword research require a browser?

Three Ways to Research Keywords From the Terminal

1. MCP Servers (AI-Assisted)

MCP servers connect AI assistants like Claude Code directly to keyword data. You ask a question in natural language, and the AI fetches and analyzes the data for you.

# In Claude Code with an SEO MCP server configured:
> "Find low-competition keywords related to 'API monitoring'"

# Claude calls the MCP server, gets data, and responds:
# Here are 15 keywords with difficulty scores and search intent...

This is the highest-leverage approach. The AI handles query expansion, intent classification, and analysis in one step.

2. Direct API Calls

If you want raw data without AI interpretation, you can hit keyword APIs directly with curl or httpie:

curl -s "https://api.delulu9.com/v1/keywords/by-keyword" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{"keyword": "api monitoring"}' | jq '.results[:5]'

This gives you structured JSON you can pipe into scripts, save to files, or process with jq.

3. CLI Wrapper Tools

Some keyword tools offer CLI wrappers that provide a familiar command-line interface:

# Example CLI tool usage
seo-keywords search "api monitoring" --intent commercial --limit 20

These are less common but useful if you want a dedicated command rather than API calls.

Setting Up MCP Keyword Research

The fastest path from zero to terminal-based keyword research:

Step 1: Install the MCP Server

npm install -g @delulu9/seo-mcp-server

Step 2: Configure Claude Code

Add to your project’s .mcp.json:

{
  "mcpServers": {
    "seo": {
      "command": "delulu9-seo-mcp",
      "args": ["--api-key", "YOUR_API_KEY"]
    }
  }
}

Step 3: Research Keywords Naturally

Now you can ask Claude Code keyword questions while you’re writing code or content:

• “What keywords should I target for a blog post about CI/CD best practices?”
• “Find keywords with commercial intent related to ‘observability tools’”
• “Compare keyword difficulty for ‘log management’ vs ‘log analytics’”

The MCP server returns data from Google, Bing, and Reddit, with intent classification and difficulty scores included.

Building Keyword Research Into Your Workflow

Content Planning in Git

Keep a keywords.md or content-plan.md in your repo. When you research keywords from the terminal, pipe results directly into your planning doc:

# Ask Claude to research and format keywords as markdown
> "Research keywords for 'kubernetes monitoring' and add them to content-plan.md as a table with difficulty and intent columns"

Pre-Commit Keyword Checks

Before publishing content, verify your target keywords are actually worth targeting:

# Quick keyword check before committing a blog post
> "Check the search volume and difficulty for the keywords in the frontmatter of content/blog/my-new-post.md"

Batch Research With Scripts

For programmatic SEO, use the API directly in your build scripts:

import requests

seeds = ["api monitoring", "log management", "error tracking"]
results = []

for seed in seeds:
    resp = requests.post(
        "https://api.delulu9.com/v1/keywords/by-keyword",
        headers={"Authorization": "Bearer YOUR_KEY"},
        json={"keyword": seed},
    )
    results.extend(resp.json()["results"])

# Filter for low-difficulty, commercial intent keywords
targets = [
    kw for kw in results
    if kw["difficulty"] == "Low" and kw["intent"] == "Commercial"
]

Why Terminal-Based Keyword Research Wins

Speed

No browser startup, no login screens, no loading dashboards. Ask a question, get an answer in seconds.

Integration

Results flow directly into your codebase. No CSV exports, no copy-paste, no context switching.

Automation

Script your keyword research. Run it in CI. Build content pipelines that start with automated keyword discovery.

Cost

Browser-based SEO tools charge $100-300/mo for their full platforms. API-first tools like Delulu9 cost $12/mo because they skip the dashboard overhead.

Comparing Terminal Keyword Research Options

When to Use Browser-Based Tools Instead

Terminal keyword research covers 80% of use cases. You still want a browser for:

• Visual SERP analysis: Seeing what actually ranks for a keyword
• Backlink research: Understanding link profiles requires visual exploration
• Competitive gap analysis: Side-by-side comparisons work better in a GUI
• Client presentations: Stakeholders expect dashboards, not terminal output

For everything else, the terminal is faster.