PocketClawvol. 1 · 2026

ZeroClaw

Privacy-first self-hosted AI agent. Local LLMs only, no cloud, AGPL-3.0.

ZeroClaw runs entirely on local hardware via Ollama or llama.cpp. Network egress is denied by default at the iptables level. No telemetry, no auto-update calls, no cloud LLM dependency. Hardware floor is real — needs 64 GB unified memory or a 24 GB GPU for usable models.

Deeper reading

Related terms

Local LLMOpenClawData residency

Found a definition that's wrong, dated or could be sharper? Email us — we update with attribution unless you'd rather we didn't.