Jump to a section

1. 1

   Install Vercel CLI globally via npm

   Requires Node.js 18+. Global flag puts the `vercel` binary on your PATH so any agent shell can invoke it.

   npm install -g vercel

2. 2

   (Alternative) Install via pnpm, yarn, or bun

   Any Node package manager works. Pick the one you already use.

   pnpm add -g vercel

3. 3

   Log in to your Vercel account

   Opens a browser tab. Pick the same sign-in method (GitHub, GitLab, Bitbucket, or email) you use on vercel.com. Token is saved to `~/.local/share/com.vercel.cli/auth.json`.

   vercel login

4. 4

   Link a local project

   Run inside your project directory. Creates a `.vercel/` folder with the project + org IDs so every subsequent command targets this project.

   vercel link

5. 5

   (Optional) Pull production env vars

   Writes `.env.local` from your Vercel project — handy for local dev and for AI agents that need the same secrets.

   vercel env pull

vercel --version

1. 1

   Install Node.js LTS (prerequisite)

   Vercel CLI is a Node package. In an elevated PowerShell, install the LTS build. Skip this step if `node --version` already returns 18+.

   winget install OpenJS.NodeJS.LTS

2. 2

   Open a FRESH PowerShell window

   The Node installer modifies PATH. Existing terminals will not see `node`/`npm` until you reopen them.

3. 3

   Install Vercel CLI globally

   Global flag is what makes `vercel` available to any shell / AI agent.

   npm install -g vercel

4. 4

   Log in to your Vercel account

   Opens your default browser. Pick the same sign-in method you use on vercel.com. The access token is stored at %APPDATA%\com.vercel.cli\auth.json.

   vercel login

5. 5

   Link a local project

   From your project directory. Select the Vercel team + project to link. Creates a `.vercel/` folder you should NOT commit.

   vercel link

6. 6

   (Optional) Pull production env vars

   Writes `.env.local` so local + agent commands see the same secrets as production.

   vercel env pull

💡 Avoid installing Node via the Microsoft Store — it sandboxes global installs and `npm install -g` silently fails. If you see EACCES errors, follow npm's permissions guide: https://docs.npmjs.com/resolving-eacces-permissions-errors.

vercel --version

1. 1

   Install Node.js 18+ via nvm (recommended)

   nvm avoids the `sudo npm -g` permissions dance and lets you switch versions per project.

   curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.7/install.sh | bash && source ~/.bashrc && nvm install --lts

2. 2

   Install Vercel CLI globally

   npm install -g vercel

3. 3

   Log in to your Vercel account

   Prints a URL + one-time code. Copy the URL to a browser (useful on headless boxes).

   vercel login

4. 4

   Link a local project

   vercel link

5. 5

   (Optional) Pull production env vars

   vercel env pull

vercel --version

1. 1

   Install via Homebrew

   Official formula. Updates with `brew upgrade gh`.

   brew install gh

2. 2

   Authenticate with GitHub

   Interactive prompts: pick GitHub.com → HTTPS → "Login with a web browser". Copy the one-time code it prints, paste it at github.com/login/device, authorize.

   gh auth login

3. 3

   Verify auth

   Should print `Logged in to github.com as <your-username>` and show the token scopes.

   gh auth status

gh --version

1. 1

   Install via winget (recommended)

   The WinGet package is the only one maintained by the GitHub CLI team. Run in PowerShell — a UAC prompt will appear.

   winget install --id GitHub.cli

2. 2

   (Alternative) Install via Scoop

   Community-maintained. Good if you already use Scoop for dev tooling.

   scoop install gh

3. 3

   (Alternative) Install via Chocolatey

   Community-maintained. Requires Chocolatey installed first.

   choco install gh

4. 4

   Open a FRESH PowerShell or Terminal window

   The installer modifies PATH — existing shells will not see `gh` until they are restarted.

5. 5

   Authenticate with GitHub

   Interactive prompts. Recommended answers: GitHub.com → HTTPS → Yes (auth git with gh credentials) → Login with a web browser. It prints a one-time code; paste it at github.com/login/device and authorize.

   gh auth login

6. 6

   (Alternative) Authenticate with a Personal Access Token

   If you cannot use the browser flow (CI, agent-only machine), create a classic PAT at github.com/settings/tokens, then pipe it in. Replace YOUR_PAT with the token.

   echo YOUR_PAT | gh auth login --with-token

7. 7

   Verify auth

   Confirms gh can talk to the API and lists the token scopes.

   gh auth status

💡 WinGet uses the official signed installer. Choco/Scoop are community forks — if anything breaks, fall back to WinGet or download the .msi from https://github.com/cli/cli/releases/latest.

gh --version

1. 1

   Add the GitHub CLI apt keyring (Debian / Ubuntu)

   For Fedora / RHEL / openSUSE, see the official install docs — each distro has a one-liner.

   (type -p wget >/dev/null || (sudo apt update && sudo apt-get install wget -y)) && sudo mkdir -p -m 755 /etc/apt/keyrings && wget -qO- https://cli.github.com/packages/githubcli-archive-keyring.gpg | sudo tee /etc/apt/keyrings/githubcli-archive-keyring.gpg > /dev/null && sudo chmod go+r /etc/apt/keyrings/githubcli-archive-keyring.gpg

2. 2

   Add the apt source list

   echo "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/githubcli-archive-keyring.gpg] https://cli.github.com/packages stable main" | sudo tee /etc/apt/sources.list.d/github-cli.list > /dev/null

3. 3

   Install gh

   sudo apt update && sudo apt install gh -y

4. 4

   Authenticate with GitHub

   GitHub.com → HTTPS → browser flow. Copy the one-time code to github.com/login/device.

   gh auth login

5. 5

   Verify auth

   gh auth status

gh --version

1. 1

   Install via Homebrew tap

   Official tap maintained by Supabase. Updates with `brew upgrade supabase`.

   brew install supabase/tap/supabase

2. 2

   Log in to Supabase

   Opens a browser, generates a personal access token, and saves it to `~/.supabase/access-token`.

   supabase login

3. 3

   Initialize a new project directory

   Creates `supabase/config.toml` and the migrations folder. Run in an empty or existing app directory.

   supabase init

4. 4

   Link to your cloud project

   Find <project-ref> in your Supabase dashboard URL: https://supabase.com/dashboard/project/<ref>. This lets `supabase db push`, functions deploy, etc. target the right project.

   supabase link --project-ref <your-project-ref>

5. 5

   (Optional) Start the local stack

   Requires Docker Desktop running. Spins up Postgres + Auth + Storage + Studio on localhost.

   supabase start

supabase --version

1. 1

   Install Scoop (if you don't have it)

   Scoop is the only supported Windows installer for Supabase CLI. Run in a NON-admin PowerShell — Scoop installs per-user.

   Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser; irm get.scoop.sh | iex

2. 2

   Add the Supabase Scoop bucket

   Pulls the bucket manifest from Supabase's official scoop repo.

   scoop bucket add supabase https://github.com/supabase/scoop-bucket.git

3. 3

   Install Supabase CLI

   Scoop downloads the prebuilt Windows binary and adds it to PATH.

   scoop install supabase

4. 4

   Log in to Supabase

   Opens your default browser. Click "Authorize" on the Supabase consent screen. The token is stored under %USERPROFILE%\.supabase\.

   supabase login

5. 5

   Initialize a new project directory

   Creates `supabase/config.toml`. Run in an empty directory (or your app's root).

   supabase init

6. 6

   Link to your cloud project

   Get <project-ref> from your Supabase dashboard URL (the 20-char slug after /project/).

   supabase link --project-ref <your-project-ref>

7. 7

   (Optional) Start the local stack

   Requires Docker Desktop for Windows installed and running.

   supabase start

💡 Supabase EXPLICITLY does not support `npm install -g supabase`. On Windows, use Scoop. Do not follow random Stack Overflow answers telling you to npm-install it globally — it will silently break.

supabase --version

1. 1

   Install via Homebrew on Linux (recommended)

   Linuxbrew works cross-distro and gets updates fastest.

   brew install supabase/tap/supabase

2. 2

   (Alternative) Install as a dev dependency

   Global npm install is NOT supported. Use it as a per-project devDependency and run via `npx supabase`.

   npm install supabase --save-dev

3. 3

   (Alternative) Download a prebuilt release

   Grab the .tar.gz for your arch from https://github.com/supabase/cli/releases and drop `supabase` into /usr/local/bin.

4. 4

   Log in to Supabase

   supabase login

5. 5

   Initialize a new project directory

   supabase init

6. 6

   Link to your cloud project

   supabase link --project-ref <your-project-ref>

7. 7

   (Optional) Start the local stack

   Requires Docker Engine running.

   supabase start

supabase --version

1. 1

   Install the signed .pkg (recommended)

   Official AWS-signed package. Most up-to-date and supported path.

   curl "https://awscli.amazonaws.com/AWSCLIV2.pkg" -o "AWSCLIV2.pkg" && sudo installer -pkg AWSCLIV2.pkg -target /

2. 2

   (Alternative) Install via Homebrew

   Community-maintained. Usually a few days behind upstream but simpler.

   brew install awscli

3. 3

   Create an IAM user + access keys in AWS Console

   Sign in to AWS Console → IAM → Users → Create user → attach a least-privilege policy (NEVER use root keys) → Security credentials tab → Create access key → select "CLI" as the use case. Copy the Access Key ID and Secret Access Key — the secret is shown only once.

4. 4

   Configure credentials

   Prompts for 4 values: Access Key ID, Secret Access Key, default region (e.g. us-east-1), default output format (json / text / table). Stored in ~/.aws/credentials and ~/.aws/config.

   aws configure

5. 5

   Test your credentials

   Returns your account ID, user ARN, and user ID. If this fails, your keys are wrong or your user has no permissions.

   aws sts get-caller-identity

aws --version

1. 1

   Install via MSI (recommended)

   Runs the official AWS-signed MSI silently. Adds `aws` to the system PATH. Open an ELEVATED PowerShell (Run as Administrator).

   msiexec.exe /i https://awscli.amazonaws.com/AWSCLIV2.msi /qn

2. 2

   (Alternative) Install via winget

   Works in a normal PowerShell — winget handles the elevation prompt.

   winget install --id Amazon.AWSCLI

3. 3

   (Alternative) Download + run the MSI manually

   If you cannot run commands as admin, download the MSI from https://awscli.amazonaws.com/AWSCLIV2.msi and double-click it.

4. 4

   Open a FRESH PowerShell window

   MSI modifies the SYSTEM PATH — open a brand-new PowerShell / Terminal so `aws` is on PATH. Existing ones will say "command not found".

5. 5

   Create an IAM user + access keys in AWS Console

   AWS Console → IAM → Users → Create user → attach a least-privilege policy (e.g. PowerUserAccess for dev, NOT root keys) → the new user → Security credentials → Create access key → select "Command Line Interface (CLI)" → confirm. Copy both the Access Key ID and the Secret Access Key. The Secret is shown only once — save it immediately.

6. 6

   Configure credentials

   Prompts for 4 values: AWS Access Key ID, AWS Secret Access Key, Default region (e.g. us-east-1), Default output (json). Stored at %USERPROFILE%\.aws\credentials and %USERPROFILE%\.aws\config.

   aws configure

7. 7

   (Optional) Set up a named profile for multiple accounts

   If you have work + personal AWS accounts, configure each under its own profile. Then use `aws --profile work s3 ls`, etc.

   aws configure --profile work

8. 8

   Test your credentials

   Returns your AWS account ID + user ARN. If this errors, your keys are wrong or the user has no permission.

   aws sts get-caller-identity

💡 NEVER commit %USERPROFILE%\.aws\credentials to git. For CI or AI agents, prefer AWS IAM Identity Center SSO (`aws configure sso`) over long-lived access keys.

aws --version

1. 1

   Download the v2 installer (x86_64)

   For ARM64 (Graviton, Raspberry Pi), use awscli-exe-linux-aarch64.zip instead.

   curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"

2. 2

   Unzip the installer

   unzip awscliv2.zip

3. 3

   Run the installer

   Installs to /usr/local/aws-cli/ and symlinks `aws` into /usr/local/bin/.

   sudo ./aws/install

4. 4

   (Alternative) Snap install

   One-liner on distros with snapd (Ubuntu, Fedora with snapd).

   sudo snap install aws-cli --classic

5. 5

   Create an IAM user + access keys

   Same as the Windows step: AWS Console → IAM → create dedicated CLI user → attach least-privilege policy → generate access keys.

6. 6

   Configure credentials

   aws configure

7. 7

   Test your credentials

   aws sts get-caller-identity

aws --version

1. 1

   Install via Homebrew (easiest)

   Pulls a signed prebuilt binary. No Rust or Node toolchain needed.

   brew install googleworkspace-cli

2. 2

   (Alternative) Install via npm

   Official npm wrapper. Requires Node.js 18+.

   npm install -g @googleworkspace/cli

3. 3

   (Alternative) Install from source with Cargo

   The CLI is written in Rust. Only needed if you want the bleeding edge.

   cargo install --git https://github.com/googleworkspace/cli --locked

4. 4

   Install gcloud CLI (for automated auth)

   Easiest auth path uses gcloud to provision the GCP project for you.

   brew install --cask google-cloud-sdk

5. 5

   Automated setup via gcloud

   Creates a Google Cloud project, enables the Workspace APIs, generates OAuth credentials, and opens a browser to log in. One command does everything.

   gws auth setup

6. 6

   Verify auth

   Should return a JSON list of your 5 most recent Drive files.

   gws drive files list --params '{"pageSize": 5}'

gws --version

1. 1

   Install Node.js LTS (prerequisite for npm path)

   Skip if `node --version` already returns 18+.

   winget install OpenJS.NodeJS.LTS

2. 2

   Open a FRESH PowerShell window

   PATH refresh. Required after Node install.

3. 3

   Install Google Workspace CLI globally via npm

   On Windows, the npm path is the easiest — no Rust toolchain, no manual PATH edits.

   npm install -g @googleworkspace/cli

4. 4

   (Alternative) Download a prebuilt Windows binary

   From https://github.com/googleworkspace/cli/releases — grab the .exe for windows-x86_64, place it in a folder on your PATH (e.g. C:\tools\gws.exe).

5. 5

   Install gcloud CLI (for automated auth)

   Google Cloud SDK includes the gcloud binary `gws auth setup` depends on. Elevated PowerShell required.

   winget install Google.CloudSDK

6. 6

   Automated setup via gcloud

   Creates a GCP project, enables the Workspace APIs (Drive, Gmail, Calendar, etc.), generates OAuth credentials, and opens a browser to complete the OAuth flow. Credentials saved to %USERPROFILE%\.config\gws\.

   gws auth setup

7. 7

   (Manual alternative) Set up OAuth credentials yourself

   If you prefer not to install gcloud: go to console.cloud.google.com/apis/credentials → Create Credentials → OAuth client ID → Desktop app. Download the JSON, save it as %USERPROFILE%\.config\gws\client_secret.json, then run `gws auth login` to complete the browser flow.

   gws auth login

8. 8

   Verify auth

   Lists your 5 most recent Drive files as JSON. If this errors with 401, re-run `gws auth setup` or `gws auth login`.

   gws drive files list --params "{\"pageSize\": 5}"

💡 The CLI is written in Rust but is distributed as prebuilt binaries — you do NOT need to install Rust or Go on Windows. Use the npm package or the GitHub release binaries.

gws --version

1. 1

   Install via Homebrew on Linux (simplest)

   Works across distros via Linuxbrew. Gets signed binaries.

   brew install googleworkspace-cli

2. 2

   (Alternative) Install via npm

   Requires Node.js 18+. Works everywhere.

   npm install -g @googleworkspace/cli

3. 3

   (Alternative) Install from source

   Requires Rust toolchain (`rustup`). Builds from the official repo.

   cargo install --git https://github.com/googleworkspace/cli --locked

4. 4

   Install gcloud CLI (for automated auth)

   Debian/Ubuntu one-liner; other distros see cloud.google.com/sdk/docs/install.

   curl https://sdk.cloud.google.com | bash && exec -l $SHELL

5. 5

   Automated setup via gcloud

   Provisions the Cloud project, enables APIs, runs the OAuth flow.

   gws auth setup

6. 6

   Verify auth

   gws drive files list --params '{"pageSize": 5}'

gws --version

1. 1

   Prerequisite: Python 3.8+

   Check with `python3 --version`. Install via Homebrew if missing.

   brew install python

2. 2

   Install huggingface_hub with CLI extras

   The `[cli]` extra installs the `hf` binary (replaces the old `huggingface-cli`).

   pip install -U "huggingface_hub[cli]"

3. 3

   (Alternative) Install via uv

   Faster if you already use uv for Python envs.

   uv tool install "huggingface_hub[cli]"

4. 4

   Log in with a user access token

   Create a token at https://huggingface.co/settings/tokens (read + write scope). `hf auth login` prompts for the token and stores it in `~/.cache/huggingface/token`.

   hf auth login

5. 5

   Verify auth

   Returns your HF username + org memberships.

   hf auth whoami

6. 6

   (Optional) Download a model

   Pulls the repo into the HF cache. Agents can reference it by repo_id afterward.

   hf download meta-llama/Llama-3.2-1B-Instruct

hf --version

1. 1

   Prerequisite: Python 3.8+

   Install via winget if missing. Tick "Add python.exe to PATH" during interactive installs.

   winget install Python.Python.3.12

2. 2

   Open a FRESH PowerShell window

   PATH refresh. Required after Python install so `pip` resolves.

3. 3

   Install huggingface_hub with CLI extras

   Installs the `hf` binary onto your Python Scripts PATH. Quotes are required in PowerShell for the bracket extra.

   pip install -U "huggingface_hub[cli]"

4. 4

   Log in with a user access token

   Generate a token at https://huggingface.co/settings/tokens (read + write scope). The command prompts for the token and stores it in %USERPROFILE%\.cache\huggingface\token.

   hf auth login

5. 5

   Verify auth

   Returns your HF username + org memberships. 401 means re-run `hf auth login`.

   hf auth whoami

6. 6

   (Optional) Download a model

   Cached under %USERPROFILE%\.cache\huggingface\hub — agents resolve by repo_id after this.

   hf download meta-llama/Llama-3.2-1B-Instruct

💡 If `hf` is not recognised after install, the Python Scripts dir is missing from PATH. Add `%USERPROFILE%\AppData\Local\Programs\Python\Python3xx\Scripts` and reopen the shell.

hf --version

1. 1

   Prerequisite: Python 3.8+ and pip

   Debian/Ubuntu example. For other distros, use your package manager.

   sudo apt update && sudo apt install -y python3 python3-pip

2. 2

   Install huggingface_hub with CLI extras

   The `[cli]` extra installs the `hf` binary onto your user bin dir (usually `~/.local/bin`).

   pip install -U "huggingface_hub[cli]"

3. 3

   (Alternative) Install via pipx or uv

   Isolated install — avoids polluting the system Python.

   pipx install "huggingface_hub[cli]"

4. 4

   Log in with a user access token

   Create a token at https://huggingface.co/settings/tokens (read + write scope). Stored in `~/.cache/huggingface/token`.

   hf auth login

5. 5

   Verify auth

   hf auth whoami

6. 6

   (Optional) Download a model

   hf download meta-llama/Llama-3.2-1B-Instruct

hf --version