{"name":"ByteWaveNetwork","version":"1.1.0","protocolVersion":"2024-11-05","description":"MCP server for ByteWaveNetwork. Exposes 14 tools: SEO/performance tools (link checking, SEO analysis, redirect tracing, page speed inspection, sitemap validation, schema markup testing, on-page SEO scoring) and AI eval runners (context retrieval, instruction following, agentic loop, thinking mode, prompt sensitivity) with get_scan_status and get_eval_result for polling async operations.","endpoint":"https://www.bytewavenetwork.com/api/mcp","tools":[{"name":"link_checker","description":"Crawl a website and find broken links, redirects, slow pages, canonical issues, and noindex pages. Returns a scan_id — use get_scan_status(tool=\"link_checker\") to poll for completion."},{"name":"seo_analyzer","description":"Crawl a website and score each page 0–100 across 22 SEO checks (title, description, headings, canonical, OG tags, JSON-LD, etc.). Returns a scan_id — use get_scan_status(tool=\"seo_analyzer\") to poll."},{"name":"redirect_tracer","description":"Follow the full redirect chain for a URL and report every hop with status codes, timing, headers, and loop detection. Returns results synchronously."},{"name":"page_speed_inspector","description":"Measure TTFB, DNS, TCP, TLS timing, HTTP version, compression, cache policy, and 12+ HTML performance checks for a URL. Returns results synchronously."},{"name":"sitemap_validator","description":"Fetch an XML sitemap (including sitemap indexes) and check every URL for broken links, noindex, and canonical mismatches. Returns a scan_id — use get_scan_status(tool=\"sitemap_validator\") to poll."},{"name":"schema_tester","description":"Extract and validate JSON-LD, Microdata, and RDFa schema markup from a single URL against 18 Schema.org types. Returns results synchronously."},{"name":"page_seo_score","description":"Run a single-page on-page SEO audit: title, description, headings, canonical, OG tags, structured data, E-E-A-T signals, content depth, and visual evidence. Returns results synchronously."},{"name":"get_scan_status","description":"Check the status of an async scan started by link_checker, seo_analyzer, sitemap_validator, schema_scan, or seo_page_checker. Returns status and summary stats. Poll every few seconds until status is \"done\" or \"complete\"."},{"name":"context_retrieval","description":"Run a needle-in-haystack long-context retrieval eval. Injects a fact into a document at configurable positions (% through the document) and scores whether the model can retrieve it. Returns an eval_id — use get_eval_result(tool=\"context_retrieval\") to poll for results."},{"name":"instruction_following","description":"Run a multi-step instruction following eval. Tests whether a model can follow complex, layered instructions across easy/medium/hard difficulty levels. Returns an eval_id — use get_eval_result(tool=\"instruction_following\") to poll for results."},{"name":"agentic_loop","description":"Run a 5-step agentic loop reliability eval using real external APIs (Wikipedia, Open-Meteo, REST Countries). Tests whether a model completes a full research chain without abandoning tools. Returns an eval_id — use get_eval_result(tool=\"agentic_loop\") to poll for results."},{"name":"thinking_mode","description":"Compare model output quality with reasoning/thinking mode ON vs OFF across math, coding, factual, and creative tasks. Returns an eval_id — use get_eval_result(tool=\"thinking_mode\") to poll for results."},{"name":"prompt_sensitivity","description":"Measure how output quality varies across 10 phrasing variants of the same task (terse to verbose, formal to informal, zero-shot to few-shot). Returns an eval_id — use get_eval_result(tool=\"prompt_sensitivity\") to poll for results."},{"name":"get_eval_result","description":"Check whether an async eval started by context_retrieval, instruction_following, agentic_loop, thinking_mode, or prompt_sensitivity has completed. Returns the full results if done, or a \"still running\" message if not yet available. Poll every 10–30 seconds."}]}