Proofsource
MCP server by Proofsource · mcp
Proofsource measures how AI answer engines (ChatGPT, Perplexity, Claude, Google AI Overviews and others) answer a brand's buyer questions: whether they mention the brand, which competitors they recommend instead, which sources they cite, and whether the fixes you ship move those numbers. Start with list_brands (every other tool takes its brand_id; it may be omitted when the workspace tracks one brand). Every data set the brand has is readable here: when unsure what data exists or which tool answers, call brand_data_catalog first. Is my brand mentioned in AI answers: get_visibility. Who does AI recommend instead of me: get_competitors. Which sources do AI answers cite: get_citations. What changed: get_changes. What to fix: list_fix_opportunities. Did my fix work: get_fix_verification. Writes: create_prompt tracks a new buyer question; mark_fix_shipped dates a fix so its effect can be verified. To change anything else the app can change: find_actions with what the user wants, then propo…
Endpoint: https://app.proofsource.co/v1/mcp
Verification: observed. Source: crawl.
Capabilities (16)
- list_brands
- list_topics
- get_visibility
- get_competitors
- get_citations
- get_changes
- list_fix_opportunities
- get_fix_verification
- get_loop
- create_prompt
- mark_fix_shipped
- add_brand
- get_scan_status
- get_report_link
- see_example