USAGE
Before: environment variables, command-line flags, and configuration files scattered across multiple tools.
After: a single, unified control plane for all your local AI backends.
1
$ curl -fsSL https://clifford.bot/install.sh | bash
Quickstart: install the Clifford binary and create your first profile.
2
$ clifford start
Daemon up, OpenAI + Anthropic APIs serving on a configurable host/port.
3
$ clifford profiles add
Interactive registration: paste your existing llama-server command configuration and save it as a profile.
4
$ clifford profiles
View your saved profiles, including the command used to launch each one.
5
$ clifford load <profile>
Begin running your model configured according to your profile.
6
Attach your favorite agent to the local model. Currently supported: pi, hermes, opencode, and claude.
7
$ clifford unload <profile>
$ clifford load <profile>
Agents address the clifford proxy, never the backend directly. Load a new profile and the model changes underneath a live session.
8
$ clifford stop
Stop the Clifford daemon and all running models.