Rankry 1.03 is live. This release focused on two things: making the platform more stable and shipping two features we’ve been working on for the past few weeks.
What’s new
AI SEO Checker
Your site might rank well on Google and still be invisible to AI models. Different crawlers, different extraction logic, different signals that matter. Traditional SEO audits don’t check any of this.
AI SEO Checker is an autonomous audit agent that crawls your site the way AI models do and generates a structured report: an AI Readiness Score from 0 to 100, a breakdown by category (crawler access, structured data, content extractability, entity clarity, FAQ structure, citation-ready blocks, freshness signals), and a prioritized list of what to fix first.
Enter a URL, get a full picture of what’s preventing ChatGPT, Claude, Gemini, and Perplexity from properly reading and citing your content.
Available on Base plan: 4 audits per month.
Custom Semantic Core
When you onboard, we auto-generate 50 buyer-intent prompts for your category. They work out of the box. But after a few weeks with the data, you start seeing gaps: audience segments that aren’t covered, languages you need to track, query patterns specific to your market.
Now you can fully customize your Semantic Core. Edit, delete, or rephrase any prompt, including the ones we generated. Add your own up to a total of 100. Replace some or all of them. Organize by custom categories. Rewrite prompts in any language for multilingual tracking.
Your Semantic Core should reflect your market, not a generic template.
Stability and performance
Beyond new features, 1.03 includes a batch of fixes under the hood:
Report generation is more stable across all five models. We resolved intermittent timeout issues that occasionally caused incomplete reports, particularly with Claude and Gemini endpoints. Retry logic has been improved so failed model calls recover gracefully without delaying the entire report cycle.
Dashboard loading is faster. We optimized how historical data is fetched and rendered, especially for accounts with several weeks of tracking history. Charts and comparison tables now load without the lag some users were experiencing.
Prompt processing pipeline has been hardened. Edge cases in prompt validation (special characters, mixed-language queries, unusually long prompts) that previously could cause silent failures are now handled properly with clear error messages.
Overall, the platform should feel snappier and more predictable. If you’ve hit any rough edges in the past few weeks, most of them are addressed in this release.
Update is live for all users. No action needed on your end.
