draw2agent 2.0.1 → 2.0.3

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -1,78 +1,78 @@
1
- # draw2agent ✏️
2
-
3
- [![npm version](https://img.shields.io/npm/v/draw2agent)](https://www.npmjs.com/package/draw2agent)
4
- [![mcp-registry](https://img.shields.io/badge/mcp--registry-io.github.zero--abd%2Fdraw2agent%402.0.0-blue)](https://registry.modelcontextprotocol.io/?q=draw2agent)
5
-
6
- Draw on your website. Your AI agent sees it.
7
-
8
- **draw2agent** is an MCP server that lets you draw annotations directly on top of your local dev page. When you submit, your IDE agent receives a screenshot, structured DOM data, and annotation context to make precise code edits.
9
-
10
- 👉 **Try it out at:** [draw2agent.vercel.app](https://draw2agent.vercel.app)
11
-
12
- ## Demo
13
-
14
- [![draw2agent demo video](https://img.youtube.com/vi/siv1ioOnOXk/maxresdefault.jpg)](https://youtu.be/siv1ioOnOXk)
15
-
16
- ## Quick Start
17
-
18
- ### 1. Add to your IDE (one-time)
19
-
20
- **Cursor** (`~/.cursor/mcp.json`):
21
- ```json
22
- {
23
- "mcpServers": {
24
- "draw2agent": {
25
- "command": "npx",
26
- "args": ["-y", "draw2agent@latest"]
27
- }
28
- }
29
- }
30
- ```
31
-
32
- ### 2. Use it
33
-
34
- Tell your agent:
35
- > "Use draw2agent to fix the navbar"
36
-
37
- 1. 🌐 Agent opens your browser with drawing tools on your page
38
- 2. ✏️ Draw circles, arrows, text directly on your website
39
- 3. 📸 Click **Submit**
40
- 4. 🤖 Agent reads the visual context and applies code changes
41
-
42
- ## How It Works
43
-
44
- ```
45
- Your Dev Page (proxied)
46
- ├── Your original page content
47
- └── Excalidraw overlay (transparent, on top)
48
- ├── Draw mode: annotate directly on the page
49
- ├── Select mode: interact with the page normally (Esc)
50
- └── Submit: screenshot + DOM + annotations → agent
51
- ```
52
-
53
- ## Tools
54
-
55
- The MCP server exposes the following tools:
56
-
57
- | Tool | Description |
58
- |---|---|
59
- | `launch_canvas` | Opens your dev page with the drawing overlay |
60
- | `launch_ipad_canvas` | Creates a tunnel and returns a QR code for remote drawing from iPad/mobile |
61
- | `launch_scratch` | Opens a standalone Excalidraw whiteboard for freehand sketching |
62
- | `get_drawing_state` | Returns screenshot, DOM nodes, and annotations for the current state |
63
-
64
- ### `launch_canvas`
65
- The core tool — proxies your localhost dev server and injects an Excalidraw overlay. Draw annotations directly on your running app, then submit to send visual context to your agent. The tool blocks until you submit.
66
-
67
- ### `launch_ipad_canvas`
68
- Same as `launch_canvas`, but exposes the proxy over the internet via a secure tunnel. Returns a QR code that you can scan from your iPad or phone to draw annotations with touch. Perfect for whiteboard-style feedback sessions.
69
-
70
- ### `launch_scratch`
71
- Opens a blank Excalidraw whiteboard — no target URL needed. Sketch UI mockups, wireframes, or diagrams from scratch. Your agent receives the drawing and implements the design.
72
-
73
- ### `get_drawing_state`
74
- Returns the last captured drawing state (screenshot, DOM nodes, annotations) without launching a new session. Useful for re-fetching context.
75
-
76
- ## License
77
-
78
- MIT
1
+ # draw2agent ✏️
2
+
3
+ [![npm version](https://img.shields.io/npm/v/draw2agent)](https://www.npmjs.com/package/draw2agent)
4
+ [![mcp-registry](https://img.shields.io/badge/mcp--registry-io.github.zero--abd%2Fdraw2agent%402.0.0-blue)](https://registry.modelcontextprotocol.io/?q=draw2agent)
5
+
6
+ Draw on your website. Your AI agent sees it.
7
+
8
+ **draw2agent** is an MCP server that lets you draw annotations directly on top of your local dev page. When you submit, your IDE agent receives a screenshot, structured DOM data, and annotation context to make precise code edits.
9
+
10
+ 👉 **Try it out at:** [draw2agent.vercel.app](https://draw2agent.vercel.app)
11
+
12
+ ## Demo
13
+
14
+ [![draw2agent demo video](https://img.youtube.com/vi/siv1ioOnOXk/maxresdefault.jpg)](https://youtu.be/siv1ioOnOXk)
15
+
16
+ ## Quick Start
17
+
18
+ ### 1. Add to your IDE (one-time)
19
+
20
+ **Cursor** (`~/.cursor/mcp.json`):
21
+ ```json
22
+ {
23
+ "mcpServers": {
24
+ "draw2agent": {
25
+ "command": "npx",
26
+ "args": ["-y", "draw2agent@latest"]
27
+ }
28
+ }
29
+ }
30
+ ```
31
+
32
+ ### 2. Use it
33
+
34
+ Tell your agent:
35
+ > "Use draw2agent to fix the navbar"
36
+
37
+ 1. 🌐 Agent opens your browser with drawing tools on your page
38
+ 2. ✏️ Draw circles, arrows, text directly on your website
39
+ 3. 📸 Click **Submit**
40
+ 4. 🤖 Agent reads the visual context and applies code changes
41
+
42
+ ## How It Works
43
+
44
+ ```
45
+ Your Dev Page (proxied)
46
+ ├── Your original page content
47
+ └── Excalidraw overlay (transparent, on top)
48
+ ├── Draw mode: annotate directly on the page
49
+ ├── Select mode: interact with the page normally (Esc)
50
+ └── Submit: screenshot + DOM + annotations → agent
51
+ ```
52
+
53
+ ## Tools
54
+
55
+ The MCP server exposes the following tools:
56
+
57
+ | Tool | Description |
58
+ |---|---|
59
+ | `launch_canvas` | Opens your dev page with the drawing overlay |
60
+ | `launch_ipad_canvas` | Creates a tunnel and returns a QR code for remote drawing from iPad/mobile |
61
+ | `launch_scratch` | Opens a standalone Excalidraw whiteboard for freehand sketching |
62
+ | `get_drawing_state` | Returns screenshot, DOM nodes, and annotations for the current state |
63
+
64
+ ### `launch_canvas`
65
+ The core tool — proxies your localhost dev server and injects an Excalidraw overlay. Draw annotations directly on your running app, then submit to send visual context to your agent. The tool blocks until you submit.
66
+
67
+ ### `launch_ipad_canvas`
68
+ Same as `launch_canvas`, but exposes the proxy over the internet via a secure tunnel. Automatically opens a new browser tab on your computer with a QR code. Scan it from your iPad or phone to draw annotations with touch. Perfect for whiteboard-style feedback sessions.
69
+
70
+ ### `launch_scratch`
71
+ Opens a blank Excalidraw whiteboard — no target URL needed. Sketch UI mockups, wireframes, or diagrams from scratch. Your agent receives the drawing and implements the design.
72
+
73
+ ### `get_drawing_state`
74
+ Returns the last captured drawing state (screenshot, DOM nodes, annotations) without launching a new session. Useful for re-fetching context.
75
+
76
+ ## License
77
+
78
+ MIT
package/dist/index.js CHANGED
@@ -14337,10 +14337,11 @@ function createMcpServer() {
14337
14337
  }
14338
14338
  const tunnelUrl = await startTunnel(proxyPort);
14339
14339
  const qr = await generateQR(tunnelUrl);
14340
- console.error(`
14341
- [draw2agent] \u{1F4F1} iPad Canvas Ready!`);
14342
- console.error(`[draw2agent] \u{1F517} Scan this QR code or open: ${tunnelUrl}`);
14343
- console.error(qr.ascii);
14340
+ const os = await import("os");
14341
+ const tempFile = path3.join(os.tmpdir(), `draw2agent-qr-${Date.now()}.html`);
14342
+ const html = `<!DOCTYPE html><html><body style="display:flex;flex-direction:column;align-items:center;justify-content:center;height:100px;min-height:100vh;margin:0;background:#1e1e2e;color:white;font-family:system-ui,sans-serif;"><h1>\u{1F4F1} Scan to Draw</h1><p style="margin-bottom:30px;opacity:0.8;">Scan this QR code from your iPad to start annotating remotely.</p><img src="${qr.dataUrl}" style="border-radius:12px;width:300px;height:300px;"/><p style="margin-top:30px;font-size:1.2rem;background:#ffffff10;padding:8px 16px;border-radius:8px;">${tunnelUrl}</p></body></html>`;
14343
+ fs3.writeFileSync(tempFile, html);
14344
+ await openBrowser(`file://${tempFile}`);
14344
14345
  clearState();
14345
14346
  const state = await waitForState();
14346
14347
  await stopTunnel();