compare(ai-app-builder): the article compares 4, the builder index lists 5

[609NFT]

Your data layer already treats VibeKit as an AI App Builder, but this article doesn't.

• data/tools/vibekit.yaml has builder_mode: true, type: web, and your own last_full_review: '2026-08-20'
• /en/tools/builder renders it, and its ItemList schema lists it at position 5 alongside Lovable, Bolt.new, v0 and Youware
• /en/tool/vibekit/ and /zh/tool/vibekit/ are both live

But this comparison still covers four. This PR closes that gap in both mirrors.

Every VibeKit value comes from your own YAML. I did not introduce new numbers: the pricing table, free-tier limits, MCP support, managed database and the 7.0 / 7.0 / 6.0 scores in Overall Ratings are the values you recorded on 2026-08-20.

What I deliberately did NOT touch:

• Generation Quality, Iteration & Editing, Technology Stack Support — those are your hands-on evaluation. I'm the maker of VibeKit, so writing my own quality prose there wouldn't be credible. Left for you if you ever want to test it.
• The title and description — they still name four. That feeds your OG images and social templates, so it's your call. Say the word and I'll update it in this PR.
• No outbound link to vibekit.bot — the other four tools are plain text here, so a self-link would stand out. Readers reach us via your existing /tool/vibekit/ page.

Honest limits are in the diff, not buried: iOS only with no Android, and the app runs in VibeKit's container rather than one you own. The zh 国内可用性 section changed from 四款 to 五款 because VibeKit needs a proxy in China too (matches china_friendly: 1 in your YAML).

npm run build passes: 335 pages, both mirrors render.

Disclosure: I work on VibeKit. You merged #78 and #103; same approach here, and same standing offer to cut anything that reads as overreach.

[vercel]

@609NFT is attempting to deploy a commit to the ww's projects Team on Vercel.

A member of the Team first needs to authorize it.