MiniMax Fleet — Multi-Instance Agent Architecture

**Designed by:** Claw (the agent who'll be using it) **Date:** Feb 16, 2026 **Status:** 🟡 Instance 1 provisioning on Vast.ai This is built from **my perspective as the consumer** — the AI agent that will route tasks to these instances. The architecture must be: 1. **Modular** — add/remove instances without touching the router 2. **Extensible** — new use cases plug in, don't require redesign 3. **Multi-instance ready** — scale from 1 to N GPUs seamlessly 4. **Self-healing** — detect down instances, failover automatically 5. **Cost-aware** — route to cheapest capable instance per task | Use Case | Context Needed | Priority | Instance Preference | |----------|---------------|----------|-------------------| | **Coding Agent** | High (full codebase) | Speed | Closest region, highest VRAM | | **Cognitive Twin** | Medium (conversation) | Quality | Any (task not latency-sensitive) | | **Office Automation** | Low-Medium | Speed | Any | | **Agentic Workflows** | High (tool chains) | Reliability | US-based (lower latency to APIs) | | **Bulk Processing** | Low | Cost | Cheapest instance | | **Quick Chat** | Low | Speed | Any available | Runs every 30 seconds per instance: - `GET /health` — is the server responding? - `POST /v1/chat/completions` — can it generate? (1-token test) - Latency measurement - GPU utilization (via Vast.ai API) - Auto-remove dead instances from routing