Namera - Name Availability Checker

Run name availability checks using the namera CLI. Always activate the venv first.

Quick Reference

source .venv/bin/activate

Run all namera commands from the project root (/Users/sidsharma/Namera).

Single name check:

namera search <name> --json

Multiple names with context (primary agent workflow):

namera find --json --context '<BusinessContext JSON>'

Rank and compare names:

namera rank --json <name1> <name2> <name3>
namera rank --json --profile fintech <name1> <name2>

Generate domain permutations:

namera compose <keyword> --common-suffixes --tlds com,io --check --json

Workflow

1. Gather Context

Ask the user for:

• What they're building (becomes description)
• Industry/niche if not obvious (becomes niche)
• Any name candidates they already have in mind (optional — you'll generate more)

2. Generate Name Candidates

This is your job as the agent. Don't just ask the user for names — generate 10-15 candidates yourself based on the business context. Include any names the user already suggested.

What research says works:

• 4-8 chars, 1-2 syllables — 37% higher recall vs longer names (HBR). Apple(5), Uber(4), Zoom(4), Stripe(6).
• CVCV patterns — alternating consonant-vowel ("Nura", "Voxly", "Zesti") maximizes pronounceability. Zero-shot test: if you hesitate saying it, customers will too.
• Sound symbolism matters — soft sounds (b, g, l, m, o) → friendly/reliable (bouba). Hard sounds (k, t, z) → precise/modern (kiki). Match the brand personality.
• Portmanteaus trending — word blends pack double meaning: "Splitly" (split+ly), "Groupon" (group+coupon). Works because they're novel yet parseable.
• Coined words find domains — invented names are far more likely to have .com available than real words.

Generation mix (aim for 10-15):

• ~5 coined CVCV words matching the brand's sound personality
• ~3 relevant word blends/portmanteaus from the business description
• ~3 evocative real words repurposed ("Harbor", "Canopy", "Forge")
• ~2 keyword + suffix ("PayHQ", "BudgetApp") as practical fallbacks
• Include any names the user already suggested

3. Build BusinessContext JSON and Run Check

{
  "description": "What the business does",
  "name_candidates": ["generated1", "generated2", "userSuggested1"],
  "niche": "fintech",
  "checks": ["domain", "trademark"],
  "scoring_profile": "default"
}

See reference.md for all fields and scoring profiles.

cd /Users/sidsharma/Namera && source .venv/bin/activate && namera find --json --context '<JSON>'

Parse the JSON output. Handle exit codes:

• 0 = success
• 1 = input error (bad JSON, no candidates)
• 2 = partial failure (some checks failed)
• 3 = all checks failed (network error)

4. Present Results

From the JSON output, highlight:

• Top available names with composite scores
• Trademark risks (any name with trademark conflicts)
• Domain availability per TLD (.com, .io, etc.)
• Recommendation with reasoning based on scores

5. Follow Up

Offer to:

• Generate another batch of candidates if nothing resonates
• Run namera rank --json --profile <profile> <names> for deeper comparison
• Run namera compose <keyword> --common-suffixes --check --json for more variations
• Check specific TLDs: namera domain <name> --tlds tech,ai,dev --json

Scoring Profiles

Use --profile <name> with namera rank:

• default -- balanced (domain 20%, trademark 20%, length 10%, pronounceability 10%)
• startup-saas -- prioritizes .com (25%), hard filter: .com must be available
• fintech -- heavy trademark weight (25%), hard filter: trademark must be clear
• consumer -- social handles (15%), pronounceability (15%)
• developer-tools -- weights .dev/.io, GitHub handle

Output Parsing

JSON output from namera find --json returns:

{
  "results": [
    {
      "check_type": "domain",
      "provider": "dns",
      "query": "name.com",
      "available": "available|taken|unknown",
      "details": {...}
    }
  ]
}

JSON output from namera rank --json returns:

{
  "ranked": [
    {"name": "x", "score": 87.0, "signals": {"domain_com": 100.0, "trademark": 100.0}}
  ]
}

Scores are 0-100 scale (100 = best).