Refora
MCP server by Refora · mcp
Refora measures how visible a brand is in AI answers. For each workspace, Refora asks a fixed set of prompts (questions a buyer would ask) to several AI models -- ChatGPT, Perplexity, Gemini, Claude, Mistral and Google AI Overviews -- on a schedule, and records which brands each answer mentions and which sources it cites. Start with list_workspaces: every other tool except get_top_cited_domains and audit_page needs a workspace_id from it. The id is for your next call; talk to the user about workspaces by name. Prompts are named by their text, which is also what get_prompt_performance takes. get_ai_answers returns the answers themselves -- what each model said, with the brands it mentions and the URLs it cites -- a page at a time. Use it to read or quote an answer; use the other tools for numbers. audit_page checks one web page's readiness for AI search (crawler access, llms.txt, structured data and more) for any public site. A workspace whose own questions are not tracked yet has…
Endpoint: https://mcp.refora.ai/mcp
Verification: observed. Source: crawl.
Capabilities (15)
- list_workspaces
- get_visibility_score
- get_share_of_voice
- get_visibility_trend
- get_competitor_comparison
- get_competitor_visibility_gap
- get_citation_sources
- get_top_cited_urls_for_topic
- get_top_cited_domains
- get_prompt_performance
- get_sentiment_breakdown
- get_ai_answers
- audit_page
- get_market_ai_presence
- get_account