Architecture: Should the embeddable bridge library support dual interactio...

Documentation for Sessions

🏗️ Architecture: Should the embeddable bridge library support dual interactio...

Date: 2026-01-28
Time: 20:54:16 UTC
Session: editor2-73.11
Agent: angular
Confidence: 95%
Status: ✅ Documented

📋 Decision Context

Question:

Should the embeddable bridge library support dual interaction methods (drag-drop AND click-to-add)?

Task: MVP 0.3 interaction method design

🎯 Decision Made

Chosen:

Yes - support both AI_ADD_TO_CANVAS (click-based, reliable fallback) and AI_DRAG_START/MOVE/END (drag-based, power users). Host adapter emits separate events: onAddToCanvas$ and onDragEnd$.

Options Considered:

Drag-drop only (too risky)
Click-to-add only (limits UX)
Both methods with separate event streams (chosen)

💭 Reasoning

Click-based 'Add to Canvas' is 100% reliable and works instantly. Drag-drop provides better UX for power users but has more complexity. Supporting both gives users choice and provides fallback if drag-drop has issues on certain devices/browsers.

📁 Files Involved

libs/libs-ai/ai-embeddable-bridge/src/lib/services/host-adapter.service.ts
libs/libs-ai/ai-embeddable-bridge/src/lib/models/bridge-events.model.ts

🔗 Related

Decision ID: dev_1769633656116_90s7m2q
Session: 73
Workspace: editor2

Auto-generated by MCP Smart Consolidation