Analyze an SEO competitor
Turn competitor evidence into decisions for the user's site. Do not create a flattering summary or instructions to copy the competitor.
Invoke Rankizz MCP tools
When using Rankizz through MCP, the initial tool list is intentionally compact. Call search_tools with a capability query or exact domain-tool name and set includeSchemas: true unless the arguments are already known. Invoke a discovered tool with call_read_tool when its readOnlyHint is true; use call_tool for every other tool. Pass the discovered name as tool and its original input object as arguments. REST callers continue to use the direct /api/v1/tools/<tool-name> endpoint.
Confirm comparison context
Identify the competitor domain, user's site, target market and language, and business question. Normalize the domain and clarify whether subdomains should be included when it matters.
Inspect provided exports, public pages, sitemaps, and previous landscape research before making paid calls.
Build an evidence stack
Use available local or public evidence for positioning, site structure, page types, authorship, structured data, internal linking, and content patterns.
Use Rankizz selectively:
get_domain_overviewfor estimated organic footprintget_ranked_keywordsfor keyword, URL, rank, intent, and traffic rowsget_serp_resultsfor priority head-to-head searchesget_backlinks_overviewfor profile summaryget_backlinks_profilefor a bounded sample of detailed referring pagesget_ahrefs_domain_ratingfor cached directional DRget_domain_keyword_suggestionswhen appropriate to the project workflow
Hosted calls require at least 10 credits. Stop Rankizz calls on a low-credit response and link to https://www.rankizz.com/billing. Use a small evidence-led sequence rather than calling every tool by default.
Check AI visibility when relevant
When the comparison includes AI answers, discover research_ai_rank_tracking_prompt with search_tools and includeSchemas: true, then use call_tool with the discovered name and arguments. Query one relevant prompt and only the requested providers: chatgpt, google_ai_mode, perplexity, gemini, or claude. Include projectId; use countryCode and languageCode for the comparison market. This is metered live research and does not save a prompt for later runs.
Inspect each provider's status. A successful tool call can contain provider errors; UPSTREAM_UNAVAILABLE means missing evidence, not zero visibility. Successful results contain a non-empty answer, brandMentioned, mentions, and citations. The citation list combines citation and source URLs; Google AI Mode may supply attached source links. Do not describe every listed source as an inline citation. A null webSearchTriggered means unknown. Preserve the prompt, provider, market, and researchedAt timestamp in the comparison.
Analyze, then compare
Read references/competitor-framework.md.
Determine:
- where visibility is concentrated
- topics, intents, and page types that drive it
- which pages appear durable versus incidental
- content, product, brand, entity, and authority advantages
- backlink patterns supported by measured data
- weaknesses, stale areas, mismatched intents, or overextended topics
- gaps relevant to the user's strengths
Distinguish observed facts, provider estimates, and inferences. Never report an estimate as first-party analytics.
Recommend differentiation
Recommend ways to serve the searcher better: clearer tools, better evidence, original data, stronger product experience, more focused intent, credible expertise, or underserved formats.
Do not reproduce copyrighted text, page structure, brand language, proprietary data, or outreach lists from the competitor.
Deliver
Return an executive comparison, evidence table, keyword and page themes, authority findings, what to learn, what not to imitate, the strongest strategic gaps, and three prioritized actions for the user's site.
Include source dates and data limitations. Recommend a competitive-landscape study when one competitor is insufficient to represent the market.
For repeated comparisons, save the reviewed prompts with add_ai_prompt,
choose engines with configure_ai_prompt_run, and explicitly request
run_ai_prompts. Reuse the same UUID runId on retries and read the stored
result with get_ai_prompt_run; there is no automatic schedule.
- Downloads
- 77All time
- Security audit
- Pass
- Last updated
- 2d ago
- Current version
- v1.0.0
- License
- MIT-0
