About this plugin
AEO and GEO analysis for WordPress
Kaldia analyzes each WordPress page in its real context to identify actionable AEO, GEO and relevant SEO improvements. It can then build and maintain structured data that accurately represents what the page actually contains.
No generic checklist. Kaldia works from the actual page.
Recommendations connect page evidence to concrete changes for search, answer systems and human readers.
When a structured-data improvement is supported by the page, Kaldia can build, preview, publish and verify the resulting JSON-LD directly from WordPress.
Free plugin. Your OpenAI API key. No Kaldia subscription.
Not another SEO score
Kaldia does not reduce a complex page to an arbitrary score from 0 to 100. It explains what can meaningfully be improved across three connected dimensions:
- Answers and content: are important answers explicit, self-contained and easy to extract?
- AEO and GEO: can answer engines identify entities, relationships, claims and useful information?
- Structured data: does the machine-readable representation describe what users can actually see?
Kaldia complements general SEO tools. It applies sound SEO principles when they improve discovery, understanding or reuse by search and answer systems.
One page. Full context. Focused recommendations.
Kaldia prepares a cleaned semantic representation of the selected page. Depending on the page, it can include:
- main content, headings and information structure;
- page type and relevant WordPress context;
- tables, lists and useful links;
- visible questions and answers;
- commerce information where available;
- existing JSON-LD;
- canonical and indexability signals.
Analysis stays constrained around AEO and GEO principles and page evidence. Recommendations explain what was observed, why it matters, what to change and how to verify it. Suggested copy follows the page language.
Kaldia does not invent studies, statistics, reviews, quotations or business facts. Missing proof becomes a research or editorial action.
From recommendation to action
Most AI audits stop after describing what could be improved. Kaldia can go further when an improvement concerns structured data.
Analyze > Understand > Recommend > Preview Schema > Apply > Verify > Undo if needed
Kaldia inspects existing JSON-LD before proposing anything. It separates what already exists, what the page proves and what can be generated safely. This helps it work alongside WordPress, WooCommerce, Yoast and other plugins without blindly duplicating their markup.
Supported improvements can be previewed, generated from visible facts, approved, published, verified on the public page, reviewed in History and undone.
Examples include FAQPage, OfferCatalog, eligible product ratings, articles, services, products, events and breadcrumb relationships when the required evidence exists.
When page content changes, Kaldia flags its analysis and machine-readable representation for review.
AI proposes. Kaldia verifies.
OpenAI helps Kaldia interpret meaning. Kaldia keeps factual control.
AI output is checked against page evidence before supported structured-data changes can be applied. The model cannot publish JSON-LD or edit page content. Kaldia builds the final graph locally from validated facts, lets an administrator preview it, then verifies the public HTML after publication.
Structured data can reduce ambiguity around entities and relationships already present on the page. It does not guarantee rankings, rich results, citations or AI visibility.
Kaldia does not copy third-party local-business ratings into self-serving rating markup.
Bring your own OpenAI API key
Kaldia has no subscription and does not resell AI credits. You connect your own OpenAI API key and choose when a page is analyzed.
Kaldia uses a centrally configured, cost-efficient model selected for structured page analysis. It sends a bounded semantic evidence package instead of blindly sending full HTML. The interface estimates analysis size and cost where possible.
OpenAI failures are reported explicitly. The API key is encrypted with WordPress secrets.
Built for useful signals, not GEO folklore
Kaldia does not promise AI rankings or citations. It does not turn every heading into a question, create arbitrary FAQs for Schema, use keyword-density recipes or impose generic content-length targets.
Every recommendation must be tied to the actual page and to a concrete AEO, GEO or supporting SEO rationale.
How it works
- Kaldia creates a lightweight local list of published content.
- You select a page and start its evidence-based analysis.
- You review recommendations and any supported Schema graph.
The initial index does not call OpenAI or crawl every public URL. Full public-page evidence is collected only for the page you choose.
Upgrading from Kaldia 1.x
Existing _kaldia_faq data remains public until a reviewed V2 graph replaces it on that page. Legacy output is excluded from V2 evidence so it cannot suppress a fresh review. Applying a V2 graph creates a verified History entry and an undo point.
Kaldia no longer injects an invisible site-wide Organization or WebSite graph.
External services
C.
com/v1/models. com/v1/responses with response storage disabled.
OpenAI processes the evidence and bills usage to the key owner’s account. This optional service is subject to:
- OpenAI Terms of Use: https://openai.com/policies/terms-of-use
- OpenAI Privacy Policy: https://openai.com/policies/privacy-policy