Allow users to define multiple Ollama endpoints (or any other locally served endpoint for that matter) , enabling seamless switching between different models or hardware configurations. Each endpoint should support:
- Local IP & Port: Direct access for low-latency local deployments.
- API URL: Standard HTTPS endpoints for remote or containerized setups.
- Nickname/Label: A user-friendly name (e.g., "My MacBook," "Workstation") to quickly identify each endpoint.
- API Toggle: An option to enable/disable endpoints based on current needs.
This flexibility is crucial for smaller dev teams running multiple agents on individual inference servers. Each agent can connect to its own locally-hosted Ollama instance, keeping inference isolated, fast, and scalable without relying on a single centralized backend.
Allow users to define multiple Ollama endpoints (or any other locally served endpoint for that matter) , enabling seamless switching between different models or hardware configurations. Each endpoint should support:
This flexibility is crucial for smaller dev teams running multiple agents on individual inference servers. Each agent can connect to its own locally-hosted Ollama instance, keeping inference isolated, fast, and scalable without relying on a single centralized backend.