🏗️ Architecture: What is the correct understanding of backend-vitto-api vs ba...

Date: 2026-02-17
Time: 21:32:20 UTC
Session: gspace1-112.5
Agent: backend
Confidence: 98%
Status: ✅ Documented

📋 Decision Context

Question:

What is the correct understanding of backend-vitto-api vs backend-ai-api responsibilities?

Task: Session 112: Clarify Vitto vs backend-ai-api architecture

🎯 Decision Made

Chosen:

backend-vitto-api = Jarvis/The Boss (orchestrator, task execution, smart routing, cost tracking, memory, skills, channels, security). backend-ai-api = Sub-agent infrastructure (agent tasks, orchestration plans, learning system, intelligence pyramid, MCP gateway, code generation, command center). Vitto delegates DOWN to backend-ai-api via connectivity services.

Options Considered:

• No alternatives documented

💭 Reasoning

Code-verified: backend-vitto-api has connectivity/backend-ai-client.service.ts that makes HTTP calls to backend-ai-api endpoints (/agents/tasks, /orchestration/plan, /learning/context, /infrastructure/*). backend-vitto-api has 4 controllers and ~50 services focused on orchestration. backend-ai-api has 47 controllers and 111+ services focused on agent management, learning, and infrastructure. The separation is: Vitto receives user requests, plans strategy, delegates to sub-agents via backend-ai-api, tracks costs, and learns from outcomes. This was initially confused in the frontend where both were treated as peers rather than boss/worker.

📁 Files Involved

• apis/backend-vitto-api/src/app/connectivity/backend-ai-client.service.ts
• apis/backend-vitto-api/src/app/engine/delegation-manager.service.ts
• apis/backend-ai-api/src/app/agents/agents.controller.ts

🔗 Related

• Decision ID: dev_1771363939996_el4jc5v
• Session: 112
• Workspace: gspace1

Auto-generated by MCP Smart Consolidation