NemoClaw
Self-hosted AI inference environment combining OpenClaw gateway with NVIDIA Nemotron-3 Super 120B model.
Links
Description
Custom setup for running AI inference locally/on own infrastructure. OpenClaw provides the gateway layer, NVIDIA Nemotron provides the model.
Download or use
Deployed on own infrastructure via Docker/k3s.
Reasoning for
Independence from cloud AI providers for specific workloads. Used in Agentic Systems architecture.
Key components:
- OpenClaw — gateway and sandbox environment (Docker k3s)
- NVIDIA Nemotron-3 Super 120B — large language model for inference
- Cloudflare Tunnel — remote access
Alternatives considered
- Ollama — simpler but less capable models
- Cloud-only (Claude API, OpenAI) — easier but vendor-dependent
Resources
Template: tool