We just launched the beta of Kintsu.ai — you connect your live WordPress site and edit it through AI-powered chat.
No page builder. No code editor. You just say "fix the mobile layout on the pricing page" and it happens.
Works with Divi, Elementor, custom themes, WooCommerce — whatever you're running. Sandbox mode lets you preview changes before they go live. One-click rollback if anything goes wrong.
Currently free to try (10 credits, no credit card). 40 beta users so far.
Would love feedback from anyone managing WordPress sites — what's the one change you keep putting off because it's too annoying to make?
Love this direction. The real friction in WordPress isn’t building — it’s making small changes quickly without breaking something. An AI layer for live edits feels like a strong unlock if execution is solid.
Interesting approach, removing friction for small live-site edits is a real pain point. How you’re handling conflicts with existing themes/plugins during edits?
Great question. We run a sandbox preview before any change goes live, so you can see exactly what the AI modified before it touches production. For theme/plugin conflicts, we analyze the site's existing stack first — if you're running Elementor with WooCommerce for example, the AI knows not to override Elementor's CSS classes or break WooCommerce hooks. If something does conflict, it flags it in the preview instead of pushing it live. That said, edge cases with heavily customized themes still come up — that's exactly what we're ironing out in beta.
Strong hook! Editing a live WordPress site through chat removes real friction.
Since this connects directly to production sites, the security model will define adoption.
A few important questions:
• Are changes executed with least-privilege permissions, or does the AI effectively get admin-level access?
• How are prompts validated before execution to prevent destructive or injected instructions?
• Is sandbox mode truly isolated from production data, or just a preview layer?
If an attacker gains access to the Kintsu account, they potentially gain control of the entire site. That makes account security, MFA, and audit logs critical.
We’re the Nautillo Pro security team, so we naturally look at privilege escalation and automation risks. We built Nautillo Pro for founders and teams who care about web app security and want to simulate real attacker paths before issues surface publicly. There’s a free version available, with DNS ownership validation required so only legitimate owners can test their own applications.
The concept is compelling. Tight execution boundaries will make site owners comfortable giving AI the keys.