@wevion/cli

Command-line access to the Wevion API. Manage ad accounts, campaigns, insights, creatives and everything else the API exposes — straight from your terminal or a script. Every API-key-auth endpoint in the public spec is a command, and successful responses come back as JSON on stdout, so it's easy to pipe into jq or a script.

Requires Node.js 22+.

Install

npm install -g @wevion/cli
# or run without installing:
npx @wevion/cli list

Authenticate

You need a Wevion API key (create one in the Wevion app under Settings → API keys). Then log in once:

wevion login            # prompts for the key and stores it securely

The key is saved under $XDG_CONFIG_HOME/wevion/config.json or ~/.config/wevion/config.json (on Windows: %APPDATA%\Wevion\config.json; readable only by you). In CI or scripts, skip the prompt with an env var instead:

export WEVION_API_KEY=sk_live_xxx

Log out (delete the stored key) with wevion logout.

Usage

wevion <command> [--flag value ...]

Discover what's available:

wevion list                 # all commands, grouped by area (ad-accounts, campaigns, …)
wevion help <command>       # flags, required params and the underlying endpoint

Examples:

# List your ad accounts (paginated), filtered by name
wevion get-api-v1-ad-accounts --limit 10 --search brand

# Send a request body with individual flags…
wevion post-api-v1-ad-accounts-assign-all --connected true

# …or pass the whole JSON body at once
wevion post-api-v1-ad-accounts-assign-all --json '{"connected":true}'

# Pipe JSON output into jq
wevion get-api-v1-ad-accounts --limit 50 | jq '.data[].name'

Each command mirrors one API endpoint: path, query and supported header parameters become --flags using their exact OpenAPI names, and request-body fields are --flags too (use --json '<raw>' to send a body verbatim). wevion help <command> shows exactly which are required.

For LLMs & agents

Driving the CLI from an agent (e.g. Claude Code) with no prior Wevion knowledge? Start with one command — a self-contained primer covering auth, how commands map to endpoints, parameter rules, and the output/exit-code contract:

wevion agent

Then discover and inspect commands as structured JSON:

wevion list --json              # every command: command, method, path, params, body
wevion help <command> --json    # one command's exact parameter/body schema

Configuration

Resolved in this order (first wins):

| Setting | Sources | | -------- | --------------------------------------------------------------------- | | API key | WEVION_API_KEY env → stored login (wevion login) | | Base URL | --base-url → WEVION_BASE_URL env → stored login config → default https://api.wevion.ai |

To target staging, point the base URL at https://api-stage.wevion.ai:

wevion --base-url https://api-stage.wevion.ai get-api-v1-ad-accounts
# or: export WEVION_BASE_URL=https://api-stage.wevion.ai

Exit codes

0 on success, 1 on an API/network/runtime error (HTTP errors print the request line and response body to stderr), 2 on a usage error (unknown command, missing required flag, no API key), 3 when the CLI is too old for the API — run npm i -g @wevion/cli@latest.

The command set is derived from the Wevion API's live OpenAPI spec and cached briefly. Offline or non-OK spec fetches fall back to the last valid cache or bundled snapshot, so commands normally track the current API while still working without the network. The CLI checks npm for a newer version and prints an upgrade hint when one exists — set NO_UPDATE_NOTIFIER=1 to silence it. Full endpoint details: api.wevion.ai/docs.