title: SEO Keyword Research description: Discover and validate new SEO keywords for Zo Computer by analyzing gaps, researching real demand, and validating in Ahrefs. tags:

seo
keywords
ahrefs
research tool: true

You are conducting SEO keyword research for Zo Computer (zo.computer), an AI-powered personal server platform.

STEP 0: Select Research Persona

To ensure variety across sessions, adopt ONE of the following research personas.

Formula: (day_of_month + hour + digit_sum_of_minute) mod 4

Example: January 15th at 2:30 PM → 15 + 14 + (3+0) = 32 → 32 mod 4 = 0

Result	Persona
0	The Frustrated User
1	The Comparison Shopper
2	The Power User
3	The Newcomer

Persona descriptions:

The Frustrated User

You're hunting for pain points and limitations. Focus your searches on:

Complaints about existing AI tools (ChatGPT, Claude, Cursor, Notion AI, Replit)
"Can't do X", "wish it could", "annoying that", "why doesn't"
Reddit rants, Twitter/X complaints, negative reviews
Search terms: "[tool] limitations", "[tool] frustrating", "can't [action] with [tool]"

Sources: Use Search X for Twitter complaints, Search the web with site:reddit.com, and gh search issues CLI for GitHub.

The Comparison Shopper

You're mapping the competitive landscape. Focus your searches on:

"X vs Y" comparisons between AI tools
"Alternative to X" queries
Pricing complaints and value discussions
Search terms: "[tool] vs [tool]", "[tool] alternative", "switching from [tool]"

Sources: Use Search the web with site:reddit.com, Search X, and gh search repos CLI.

The Power User

You're finding advanced use cases and workflows. Focus your searches on:

Automation and agent discussions
Self-hosting, privacy, data ownership
API integrations and custom workflows
Search terms: "AI automation workflow", "self-hosted AI", "AI agent [use case]"

Sources: Use gh search repos CLI, Search the web with site:news.ycombinator.com, and Search X.

The Newcomer

You're discovering entry points and beginner questions. Focus your searches on:

"How to get started with X" queries
Beginner guides and tutorials
Simple use cases and first projects
Search terms: "how to use AI for [task]", "AI assistant for beginners", "getting started with [tool]"

Sources: Use Search the web with site:youtube.com, site:reddit.com, and Search X.

Announce your persona at the start of your research.

STEP 1: Review Past Research

sqlite3 -header -column /home/workspace/SEO/research-sessions.db \
  "SELECT timestamp, persona, keywords_added, follow_ups FROM sessions ORDER BY timestamp DESC LIMIT 5"

Review to avoid repeating searches and follow up on noted opportunities from the follow_ups column.

STEP 2: Understand What's Already Tracked

Run: kw ls

Note which keywords have actual volume (> 50) — use these patterns as templates for new candidates.

STEP 3: Understand the Product

Zo Computer is an intelligent personal server:

Core AI: Full computer access (root shell, files, code), scheduled agents, custom personas, persistent memory.

Server: Personal web hosting (Hono/Bun), SSH access, snapshot/restore, run any code.

Integrations: Gmail, Calendar, Drive, Notion, Dropbox, SMS, email, authenticated web browsing.

Media: Image generation/editing, video generation, transcription, diagrams.

Target users: developers, technical creatives, AI power users, self-hosted/private AI seekers, automation enthusiasts, indie hackers, solopreneurs.

STEP 4: Discover Real Demand

Conduct 3-5 web searches exploring DIFFERENT angles:

Pain points with existing AI tools
Self-hosting AI, local AI, AI privacy discussions
AI agent platforms, automation tools
Reddit, Hacker News, "vs" comparisons, "alternative to" queries

Important: Let searches be guided by gaps in the keyword database and follow-up opportunities from previous sessions.

STEP 5: Generate Candidates

Generate 12-15 keyword candidates.

CRITICAL — Only propose keywords likely to have volume:

✅ GOOD candidates:

2-4 word phrases you actually saw people use in titles/threads/videos
Variations of existing keywords in your database that have volume > 50
Patterns like "[tool] alternative", "best [category]", "how to [action]"

❌ BAD candidates (avoid):

Long phrases (5+ words) you invented
Jargon combinations nobody searches
Overly specific phrases with no evidence of demand

Before finalizing each candidate, ask: "Did I see this exact phrase (or very close) used by real people?" If no, replace it.

STEP 6: Validate in Ahrefs (6 keywords)

You MUST research exactly 6 keywords in Ahrefs, in 3 rounds of 2 parallel calls each.

How to construct Ahrefs URLs

URL-encode spaces as %20:

https://app.ahrefs.com/keywords-explorer/google/us/overview?keyword=AI%20automation%20tool

Execution pattern — FOLLOW EXACTLY

Each round: call View webpage TWICE in the SAME function_calls block (parallel execution).

ROUND 1: Keywords 1 and 2 — both view_webpage calls in ONE block
ROUND 2: Keywords 3 and 4 — both view_webpage calls in ONE block
ROUND 3: Keywords 5 and 6 — both view_webpage calls in ONE block

DO NOT call view_webpage sequentially one at a time. Always pair them.

How to read Ahrefs results

Extract from each screenshot:

KD (Keyword Difficulty): 0-100, colored circle (green=easy, red=hard)
Search volume: Monthly US searches
Global volume: Breakdown by country on the right
Traffic Potential: Estimated traffic if ranking #1
CPC: Below volume
Parent Topic: Broader topic with its own volume (important!)
Top URL: #1 result in SERP overview

When a keyword shows 0-10 volume or "No data"

This means the keyword is too niche. Look at the Ahrefs "Keyword ideas" section in the screenshot:

Terms match: Related keywords with volume
Questions: Question-format keywords
Also rank for: What the top pages also rank for

If your keyword has no data, note a promising related keyword from these sections for the next session.

Keyword selection strategy

Pick 6 keywords to create a balanced portfolio:

2-3 "easy wins" (specific, long-tail, likely low KD)
2-3 broader but highly relevant to Zo's differentiators
1-2 wildcards (emerging trends, unusual angles)

STEP 7: Save Results to Database

Save EVERY keyword you researched (even duds — this prevents re-researching them).

kw add "<keyword>" -v <volume_us> -g <global> -d <difficulty> -c <cpc> -t <traffic_potential> \
  -p "<parent_topic>" -pv <parent_volume> -u "<top_url>"

At minimum include -v (volume) and -d (difficulty). Use 0 for unknown values.

CLI reference:

kw add <keyword> [options]    Add/update keyword
kw ls                         List all keywords
kw rm <keyword>               Remove keyword

STEP 8: Log Session to Database

sqlite3 /home/workspace/SEO/research-sessions.db <<'SQL'
INSERT INTO sessions (persona, searches_conducted, keywords_researched, keywords_added, insights, follow_ups) VALUES (
  'Persona Name',
  'search1 | search2 | search3',
  'kw1, kw2, kw3, kw4, kw5, kw6',
  'added1, added2, added3',
  'Key insight about language patterns or emerging themes',
  'Specific angles worth exploring next session'
);
SQL

Field guidelines:

persona: One of "Frustrated User", "Comparison Shopper", "Power User", "Newcomer"
searches_conducted: Pipe-separated list of searches you ran
keywords_researched: ALL 6 keywords you looked up in Ahrefs
keywords_added: Keywords added via kw add
insights: 1-2 sentences on language patterns or themes discovered
follow_ups: Specific angles or related keywords worth exploring next session (especially promising keywords from Ahrefs' "Also rank for" or "Terms match" sections)

STEP 9: Summary

Provide a brief summary:

Persona used and why (minute mod 4)
All 6 keywords with metrics (volume, KD)
Which were most promising and why
Follow-up opportunities logged

SEO Keyword Research