Skip to content

Editing with AI

Open AI Chat from the left rail (the ✦ icon), or pop it into its own window with the ↗ button. You’ll need to be signed in.

Type what you want and press Enter. The agent reads and edits your project’s files directly — you’ll see its progress (thinking, file reads, edits) stream in, and the canvas refreshes as it works.

  • Enter sends · Shift+Enter newline · ⌘Enter queues a follow-up while a run is still going.
  • stop aborts the current run.

Drop an image onto the chat composer, or paste one (⌘V — handy for screenshots). Thumbnails appear above the text box; remove any with ✕.

PNG, JPEG, WebP, and GIF are supported (up to six images per message, 4MB each). Pick a vision-capable model in the model dropdown when the image matters — not every model can see attachments.

The fastest way to talk about something specific: click it on the canvas, then hit Ask AI in the Inspector. The chat opens with that element attached — a small chip shows the tag and source location — and the agent works on exactly that spot. Remove the chip with ✕ if you change your mind.

Give the AI lasting context — voice, design notes, project plans — from the Config panel’s Docs tab in the left rail (⚙ icon):

Kind Where it lives What the AI does with it
Rules .plyy/rules/*.md Always included in the prompt (keep these short).
Docs .plyy/docs/*.md Listed for the AI to read when relevant (styleguides, design.md, plans).
  • Drop a .md file onto the panel (or onto the window — markdown goes to Docs, not Assets).
  • Click a file to read and edit it in the panel (⌘S saves), or hit Code to open it in the code editor.
  • Attach (or @ on a row) pins that doc to your next chat message so the AI reads it before editing.

Optional frontmatter on a file: apply: always or apply: optional.

Two dropdowns under the composer:

  • Modedefault edits anything; plan is read-only (ask questions, get a plan, nothing changes); web focuses on your site’s pages and components; cms focuses on the CMS side.
  • Model — pick any model, or leave auto. Bigger models are smarter and slower; all of them draw from the same AI credit pool.

The tab strip above the transcript holds up to 10 chats. + starts a new one; each runs independently — start a long task in one tab and keep working in another. A pulsing dot marks tabs with a run in flight, and transcripts catch up when you switch back. Chats survive app restarts. Close a tab with ✕ (finished chats only).

Edits show up as diffs under each response. If a run didn’t go the way you wanted, revert these edits puts every touched file back.

After a run that changed files, Plyy watches the dev server and the page console for a moment. If the change broke something, you’ll see an amber auto-verify follow-up where the agent fixes its own mistake — at most twice per prompt, and never when you have follow-ups queued.