The ThoughtForge MCP server
ThoughtForge exposes its measure-and-fix loop as a Model Context Protocol (MCP) server, so Claude, ChatGPT and any other MCP client can run it directly in a conversation.
Eleven tools: list tracked brands, run a Brand Health scan, read scores and top fixes, list the leakage points where a model swapped your brand for a competitor, list articles, read an article's GEO gaps, rewrite an article for citability, generate a comparison table, generate structured FAQs, draft counter-content briefs, and schedule a LinkedIn post.
Most AI visibility MCP servers are read-only reporting. This one executes the fix: the rewrite, the FAQ block and the comparison table are written back to the article.
Authentication is OAuth 2.1 with dynamic client registration, and every tool runs as the signed-in ThoughtForge user, scoped by row-level security.
New accounts get one full brand scan free, so you can see the findings before subscribing. Applying fixes and re-scanning on a schedule are on the paid plans.
Endpoint: https://hzzvddioivvtwkjjqavq.supabase.co/functions/v1/mcp — add it as a custom connector with Streamable HTTP transport.