🏗️ Architecture: How should Vitto's chat detect and adapt to different user l...

Date: 2026-02-11
Time: 19:30:08 UTC
Session: gspace1-105.23
Agent: backend
Confidence: 90%
Status: ✅ Documented

📋 Decision Context

Question:

How should Vitto's chat detect and adapt to different user layers (developer/platform_builder/end_user)?

Task: Session 105.5: Vitto layer-aware chat and Mind integration

🎯 Decision Made

Chosen:

Layer detection via auth context + URL context, passed as metadata in chat requests. VittoChatService sends layer to backend, backend adapts system prompt, tool availability, and response style per layer.

Options Considered:

• No alternatives documented

💭 Reasoning

Vitto already has VittoLayer type ('developer' | 'platform_builder' | 'end_user'). The /ai-api/chat endpoint accepts context. Rather than duplicating layer detection client-side, pass the layer from the frontend (which knows the app/route/user role) and let the backend adapt. This keeps the frontend thin and the intelligence server-side.

📁 Files Involved

• libs/libs-ai/gograb-ai-lib/src/lib/services/vitto-chat.service.ts
• apis/backend-vitto-api/src/app/types/vitto.types.ts

🔗 Related

• Decision ID: dev_1770838208270_sv3elh5
• Session: 105
• Workspace: gspace1

Auto-generated by MCP Smart Consolidation