oc-local

To run tasks on a local Ollama model, you direct the task to execute locally by selecting an available Ollama model. This enforces local-only operation, embedding all context into prompts for secure, offline processing without cloud APIs.

Yes, you can use a remote Ollama server for agentic workflows. The local model execution supports flexible modes, allowing optional remote Ollama endpoints while guiding model setup and mode switching for both single-shot and agentic tasks.

You can choose and save a default model in Ollama during the model selection process. The setup guides you through picking from available Ollama models to execute tasks, ensuring your preference is saved for future local processing.

Yes, local model execution requires embedding all context into prompts. Because the task processes on a local Ollama model instead of Claude, you must include all necessary information directly within the prompt to ensure accurate offline processing.

The best way to avoid cloud APIs for private LLM processing is executing tasks on a local Ollama model. This approach protects data privacy, reduces latency, and supports both one-shot tasks and agentic workflows securely.

You should use a local Ollama model for single-shot tasks to enable secure, offline processing and avoid cloud APIs. This method protects data privacy, reduces latency, and allows flexible model selection for your specific workflow.