Advanced
Local Claude Code server
Since version 1.0.13, Oriloq exposes a server compatible with the Claude Code CLI on your machine. You can use the local model or provider configured in Oriloq (Gemma, Ollama, OpenAI-compatible providers) as a reasoning backend for Claude Code, without depending on the Claude models hosted by Anthropic.
What is it for?
Claude Code is a CLI designed to assist software development. By default, it relies on the Claude models hosted by Anthropic. The server exposed by Oriloq offers an alternative: the CLI talks to Oriloq via the compatible API, and the model or provider configured in Oriloq (local Gemma, Ollama, OpenAI-compatible provider) processes the requests.
- →No cloud inference cost.
- →No network dependency, works offline.
- →Your sources and prompts never leave your machine.
Activation
The server starts from the dedicated screen in Oriloq. Once launched, the local URL and port are displayed. Then configure your Claude Code CLI to point at this URL in order to route requests to Oriloq.
Typical use cases
- •Development on proprietary or NDA-covered code.
- •Working on the move without a reliable connection.
- •Exploration without quota or cost constraints.
Limits
Performance and quality depend on the model or provider selected in Oriloq and your Mac's resources. For very complex tasks, cloud models are generally more accurate. The local server is primarily an alternative for confidentiality, cost and offline availability.