mentions-check#

Focused AI brand-visibility skill. Use only when the user explicitly asks whether a brand shows up in AI answers — its mention, rank, or citations for a specific question — not when a brand name merely appears in passing.

When to use this skill#

DO use when the user asks:

• "Is HubSpot mentioned when people ask ChatGPT about CRMs?"
• "Does Perplexity cite Stripe for 'best payment API'?"
• "What rank does Notion get in AI answers for 'best note-taking app'?"
• "Show up in Google AI Overviews for 'project management software'?"
• "Which brands does ChatGPT name first for 'top observability tools'?"

Do NOT use when:

• A brand or query appears incidentally in context
• The user is discussing GEO/AI search abstractly without a specific brand + query to measure
• The user has not signaled they want a visibility check

Each call costs credits (from $0.02). When the intent is ambiguous, confirm the brand and query with the user before calling.

Tools#

check_mentions — from $0.02 per call#

Ask one question against one or more AI surfaces and see whether brand is mentioned, where it ranks, and what got cited.

Arguments:

• query (required) — the natural-language question to test, e.g. "best CRM for startups"

• brand (required) — the brand name to look for, e.g. "HubSpot"

• mode — which surface(s) to hit (default quick):

  • quick ($0.02) — all 4 LLM APIs in parallel (ChatGPT, Claude, Gemini, Perplexity)
  • ai_overview ($0.05) · bing_copilot ($0.05)
  • chatgpt_live ($0.10) · gemini_live ($0.10) · ai_mode ($0.10)
  • perplexity_live ($0.25) — live Perplexity UI scrape with citations + fan-out
  • all_live ($0.50) — all 6 live UI surfaces in one call

• providers — subset of ["chatgpt","claude","gemini","perplexity"] (only applies to mode:quick)

• runs — integer 1–7, repeat the query N times to measure volatility (multiplies cost)

• region — locale hint, e.g. "us", "gb"

Example response ( mode:quick):

json

{
  "id": "chk_...",
  "mode": "quick",
  "query": "best CRM for startups",
  "brand": "HubSpot",
  "providers": {
    "chatgpt":    { "mentioned": true,  "rank": 1,    "context": "...", "citations": [{ "url": "https://...", "title": "..." }], "fan_out": [] },
    "claude":     { "mentioned": true,  "rank": 2,    "context": "...", "citations": [], "fan_out": [] },
    "gemini":     { "mentioned": false, "rank": null, "context": "",    "citations": [], "fan_out": [] },
    "perplexity": { "mentioned": true,  "rank": 3,    "context": "...", "citations": [{ "url": "https://...", "title": "..." }], "fan_out": [] }
  },
  "cost_cents": 2,
  "balance_after_cents": 98,
  "cache_hit": false
}

Authentication#

Set MENTIONSAPI_KEY to your MentionsAPI key (format lvk_live_...). Free key with $1 credit at https://mentionsapi.com/signup — no card.

Pricing#

Pay-as-you-go. No subscriptions, no monthly tiers, credits never expire. $1 free signup credit, $5 minimum top-up. Per-call cost is set by mode (see the table above). Failed calls are not billed.

Errors#

Every call returns a Python dict. On success it's the API response; on failure it has an error key:

• {"error": "auth_required", ...} — MENTIONSAPI_KEY not set (includes signup URL)
• {"error": "auth_invalid", ...} — key rejected; rotate at https://mentionsapi.com/app/keys
• {"error": "insufficient_balance", ...} — wallet empty; top up at https://mentionsapi.com/app/billing
• {"error": "rate_limit_exceeded", ...} — back off and retry
• {"error": "mode_roadmap", ...} — deep/ change_track are on the roadmap; use a shippable mode

API reference#

• Endpoint reference: https://mentionsapi.com/docs/api/check
• OpenAPI spec: https://mentionsapi.com/openapi.json
• Hosted MCP server (alternative to this skill): npx @mentionsapi/mcp

Independence and trademarks#

MentionsAPI is an independent service and is not affiliated with, endorsed by, or sponsored by OpenAI, Anthropic, Google, Microsoft, or Perplexity. All product names are the property of their respective owners.