Running a large language model on your own hardware instead of calling a cloud API, giving you full data privacy and zero per-token costs.
A local LLM runs on your machine or private server. Your data never leaves your infrastructure. There are no per-token API costs — just hardware and electricity.
OpenClaw supports local LLMs through Ollama integration. Configure Ollama as your LLM provider, point it at your local model, and the agent uses it just like any cloud provider. Same MCP tools, same channels, same security controls.
Local models are smaller than cloud models (7B-70B parameters vs 100B+), so they trade some intelligence for privacy and cost savings. For many tasks — customer FAQ, document search, basic reasoning — local models perform well.
Some data cannot leave your network. Healthcare records, legal documents, financial data — if cloud API calls violate your data policy, local LLMs are the answer. They also eliminate per-token costs for high-volume use cases.
Clawctl supports Ollama as a first-class LLM provider. Configure it in the setup wizard just like Anthropic or OpenAI. All Clawctl security features — audit trails, approval workflows, egress filtering — work identically with local models.
Try Clawctl — 60 Second DeployAny model supported by Ollama — Llama, Mistral, Phi, Gemma, and more. Check Ollama's model library for the full list.
Minimum 16GB RAM for 7B models. 32GB+ for 13B models. GPU recommended but not required.
Yes. Use local models for routine tasks and cloud models for complex reasoning in a multi-agent setup.
BYOK (Bring Your Own Key)
A model where you provide your own LLM API key (Anthropic, OpenAI, etc.) instead of the platform providing one, giving you full cost control and model choice.
Model Routing
Directing different agent tasks to different LLM models based on complexity, cost, or speed requirements.
Cost Optimization
Strategies for reducing LLM and infrastructure costs when running AI agents without sacrificing quality or reliability.
AI Agent Runtime
The execution environment that hosts an AI agent, managing its lifecycle, tool access, memory, and communication with LLM providers.