Prompt Annotations

When vibe coding, your prompts are artifacts too.

About the plugin

Prompt Annotations is a Figma plugin that captures your vibe coding chat transcript and places a structured annotation frame next to your design in the file. It keeps the prompts, AI summary, and full back and forth with your design so your whole team has the context behind the work.

The Problem

Every time you finish a vibe coding session and move the work into Figma, the same thing happens. The design makes it. The reasoning doesn't.

The prompts that shaped the UI, the constraints you gave the AI, the decisions made along the way. Gone when the tab closes. What lands on the canvas is an artifact without a memory. Nobody else on the team knows what was asked, what was tried, or what got left out.

How it works

  1. Finish your session in Cursor, V0, Bolt, Lovable, or Claude.ai

  2. Paste or upload your transcript into the plugin

  3. A structured annotation frame appears next to your design — summary, transcript, timestamp, and source tool included

An API key for Claude or ChatGPT enables an AI-generated summary of what was built. The plugin works without one.

Supported tools

Cursor · V0 · Bolt · Lovable · Claude.ai · Terminal

The flow

Step by step
01
🧑‍💻
Vibe code your design
Build your UI in any vibe coding tool. Select the matching source tool in the plugin.
Cursor V0 Bolt Lovable Claude.ai Terminal
02
📋
Capture the transcript
Export from Cursor's command palette as a .md file, or Cmd+A, Cmd+C from the chat panel in any web tool.
03
🔍
Smart parsing per tool
The plugin knows each tool's format. It strips timestamps, loading states, and chrome — keeping only the real conversation.
04
Generate Annotation
Click Generate. API key is optional.
With key — AI writes a 2–3 sentence summary
No key — placed with placeholder text
05
📐
Annotation on canvas
Placed 30px right of your selected frame, top-aligned.
Session Title
editable
What Was Built
AI summary
Chat Transcript
turns
Footer
timestamp
06
🤝
Your whole team gets context
Lives in the Figma file — where everyone already works.
Designers understand previous sessions
Engineers see intent, not just spec
Product review what was asked of the AI

Who it's for

Designers who want their team to understand the work, not just see it. Engineers who need context during handoff, not just a spec. Product managers who want to know what was actually asked of the AI.

The annotation lives in the Figma file — where your whole team already works.

Ready to keep your prompts with your designs?