Tools
visibility_check
POST /api/v1/visibility · MCP: visibility_check · priced per model asked
| Parameter | Type | Notes |
|---|---|---|
brand * | string (1–100) | brand to look for; name variants are recognised |
question * | string (5–500) | the customer's question |
models | string[] (max 5) | e.g. ["chatgpt","gemini"]; default: all available models |
competitors | string[] (max 10) | competitors to track |
brand_domain | string | brand domain, to detect if it is cited as a source |
Per model: brand.mentioned, brand.mention_count, brand.list_rank, brand.first_mention_percent, excerpt, competitors[], brand_domain_cited, sources[], answer. Summary: models_asked, models_answered, mentioned_in, visibility_percent. Models that fail are refunded.
geo_audit
POST /api/v1/geo-audit · MCP: geo_audit · priced per call
Input: url * (public domain or URL). Returns score (0–100), checks[] (id, status pass/warn/fail/info, weight, params), bots[] (18 AI crawlers, allowed or blocked by robots.txt or firewall) and llms (llms.txt and llms-full.txt). Check ids: https, robots_txt, ai_search_bots, ai_firewall, ai_training_bots, llms_txt, llms_links, llms_full_txt, sitemap, sitemap_in_robots, sitemap_fresh, js_rendering, snippet_directives, org_schema, website_schema, trust_pages, open_graph, html_lang. Unreachable sites are refunded; private or internal addresses are rejected without charge.
Response format
{ "request_id": "req_…", "tool": "visibility_check", "gross_cents": 70, "charged_cents": 35, "refunded_cents": 35, "balance_cents": 1930, "result": { } }
gross_cents is reserved before the call, charged_cents is what was actually paid. balance_cents is a snapshot at response time.