Quickstart
1) Install
bash
brew tap tappunk/tap
brew trust tappunk/tap # required on recent Homebrew versions
brew install tappunk/tap/tnk2) Initialize configuration
bash
tnk init
tnk config init
tnk config showtnk initpopulates~/.config/tnk/from tnk-specs.tnk config initcreates~/.config/tnk/tnk.tomlif missing.
Set default_model in ~/.config/tnk/tnk.toml to the model name your host inference server serves:
toml
default_model = "ai-fast"3) Start your inference server
tnk does not manage the engine. Run an OpenAI-compatible server on the host at the configured server_port (default 9931), then verify:
bash
curl -s http://127.0.0.1:9931/health4) Start a project sandbox
bash
cd ~/code/myproject
tnk run # boots the VM and provisions the default profile
tnk sandbox shelltnk sandbox shell is also available on its own. For quick commands without a full interactive shell:
bash
tnk sandbox shell --no-tty -c "echo hello"5) Check status and tear down
bash
tnk sandbox ls # list sandboxes
tnk doctor # health checks
tnk shutdown # stop all sandboxes