{"id":33989,"date":"2025-05-22T12:27:11","date_gmt":"2025-05-22T10:27:11","guid":{"rendered":"https:\/\/www.graviton.at\/letterswaplibrary\/an-alternative-cloudflare-autorag-mcp-server\/"},"modified":"2025-05-22T12:27:11","modified_gmt":"2025-05-22T10:27:11","slug":"an-alternative-cloudflare-autorag-mcp-server","status":"publish","type":"post","link":"https:\/\/www.graviton.at\/letterswaplibrary\/an-alternative-cloudflare-autorag-mcp-server\/","title":{"rendered":"An Alternative Cloudflare AutoRAG MCP Server"},"content":{"rendered":"<p><!-- SC_OFF --><\/p>\n<div class=\"md\">\n<p>I built an MCP server that works a little differently than the Cloudflare AutoRAG MCP server. It offers control over match threshold and max results. It also doesn&#8217;t provide an AI generated answer but rather a basic search or an ai ranked search. My logic was that if you&#8217;re using AutoRAG through an MCP server you are already using your LLM of choice and you might prefer to let your own LLM generate the response based on the chunks rather than the Cloudflare LLM, especially since in Claude Desktop you have access to larger more powerful models than what you can run in Cloudflare.<\/p>\n<\/div>\n<p><!-- SC_ON -->   submitted by   <a href=\"https:\/\/www.reddit.com\/user\/brass_monkey888\"> \/u\/brass_monkey888 <\/a> <br \/> <span><a href=\"https:\/\/github.com\/noops888\/cf-autorag-mcp\">[link]<\/a><\/span>   <span><a href=\"https:\/\/www.reddit.com\/r\/datasets\/comments\/1ksmjny\/an_alternative_cloudflare_autorag_mcp_server\/\">[comments]<\/a><\/span><\/p><div class='watch-action'><div class='watch-position align-right'><div class='action-like'><a class='lbg-style1 like-33989 jlk' href='javascript:void(0)' data-task='like' data-post_id='33989' data-nonce='65e0e39b87' rel='nofollow'><img class='wti-pixel' src='https:\/\/www.graviton.at\/letterswaplibrary\/wp-content\/plugins\/wti-like-post\/images\/pixel.gif' title='Like' \/><span class='lc-33989 lc'>0<\/span><\/a><\/div><\/div> <div class='status-33989 status align-right'><\/div><\/div><div class='wti-clear'><\/div>","protected":false},"excerpt":{"rendered":"<p>I built an MCP server that works a little differently than the Cloudflare AutoRAG MCP server. It&#8230;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[85],"tags":[],"class_list":["post-33989","post","type-post","status-publish","format-standard","hentry","category-datatards","wpcat-85-id"],"_links":{"self":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts\/33989","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/comments?post=33989"}],"version-history":[{"count":0,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts\/33989\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/media?parent=33989"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/categories?post=33989"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/tags?post=33989"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}