EdgeAI CLI

Customer lifecycle CLI for EdgeAI. The repository is cli; the installed binary is edgeai.

Installation

Recommended:

pipx install edgeai-cli
edgeai login
edgeai install

Alternatives:

uv tool install edgeai-cli
python3 -m pip install --user edgeai-cli

While this repo is private, GitHub Release binaries are available to repo users. PyPI is the public customer distribution path.

TestPyPI installs

After publishing a test release to TestPyPI, install the exact test version with one of these commands:

uv tool install --default-index https://test.pypi.org/simple/ edgeai-cli==0.1.0

pipx install --pip-args="--index-url https://test.pypi.org/simple/" edgeai-cli==0.1.0

python3 -m pip install --index-url https://test.pypi.org/simple/ edgeai-cli==0.1.0

For a throwaway virtual environment with uv:

uv venv /tmp/edgeai-test
source /tmp/edgeai-test/bin/activate
uv pip install --default-index https://test.pypi.org/simple/ edgeai-cli==0.1.0
edgeai --help

Commands

edgeai login
edgeai login --api-key <key>
edgeai login --no-browser
edgeai whoami
edgeai device register --name shop-floor
edgeai install
edgeai install tts stt agents-sdk friday
edgeai update
edgeai update tts
edgeai status
edgeai doctor

edgeai install with no product arguments installs every product the API key can access on the current platform. Explicit product arguments narrow the install or update to tts, stt, agents-sdk, and/or friday.

Local state defaults to ~/.edgeai and can be overridden with EDGEAI_DIR. The API base URL defaults to https://runedge.ai/api and can be overridden with EDGEAI_API_BASE_URL.

Installed product state is stored in ~/.edgeai/products.json. Product tarballs and models are downloaded into ~/.edgeai/downloads, extracted under ~/.edgeai/<product>, and launchers are generated in ~/.edgeai/bin.

Shell completion

Completion scripts are generated on demand and are intentionally opt-in:

edgeai completion zsh
edgeai completion bash
edgeai completion fish

Install the generated script using your shell's preferred completion location. The CLI does not edit shell profiles automatically.

Development

go mod tidy
go test ./...
go run ./cmd/edgeai --help

Releases

Release outputs are built by .github/workflows/release.yml:

• Native archives for darwin-arm64, linux-amd64, and linux-arm64
• checksums.txt
• PyPI wheels for the edgeai-cli package, installing the edgeai command

Build artifacts without publishing:

gh workflow run release.yml \
  --ref main \
  -f version=0.1.0 \
  -f publish_target=none

Publish a test release to TestPyPI:

gh workflow run release.yml \
  --ref main \
  -f version=0.1.0 \
  -f publish_target=testpypi

Then verify the TestPyPI package:

uv tool install --default-index https://test.pypi.org/simple/ edgeai-cli==0.1.0
edgeai --help

Then tag and push the release:

git tag v0.1.0
git push origin v0.1.0

PyPI Trusted Publishing must be configured for the edgeai-cli project with:

• Repository: RunEdgeAI/cli
• Workflow: release.yml
• Environment: pypi

Configure TestPyPI the same way with environment testpypi.