CLI

The w6w command gives you w6w from a terminal: check your access, list what you can run, run it, and manage workflows, documents and vars. It prints tables for people, raw JSON with --json for scripts, and exit codes your CI can act on.

Before you start

Node.js 18 or later.
A w6w base URL and an API token. See Configure.

1. Install the CLI

npm install -g @w6w/cli
w6w --version

You should see a version number.

2. Set your credentials

export W6W_BASE_URL=https://<your-host>   # your API origin, no trailing path
export W6W_TOKEN=…                         # Studio → Settings → API tokens

You can also pass --base-url and --token on any command, and they win over the environment. Prefer the environment variable for the token, so it stays out of your shell history.

3. Check your access

w6w me

You should see your tenant, account, user id and role, followed by the CLI’s version. w6w info does the same thing.

4. Find something to run

w6w connections list
w6w workflows list
w6w functions list

Each prints a table with ids: conn_… for connections, wf_… for workflows, fn_… for Functions. Shorter forms work too: conn for connections, wf for workflows, docs for documents, and ls for list, so w6w conn ls lists your connections.

5. Run an action

w6w run conn_01H… --action send_message --payload '{"channel":"#general","text":"Hello from w6w"}'

You should see the result’s kind, action, then the action’s output as JSON. --payload takes a JSON object, so wrap it in single quotes in your shell.

If you have exactly one connection for an app, name the app instead of the connection id:

w6w run --app slack --action send_message --payload '{"channel":"#general","text":"Hi"}'

If you have several connections for that app, the command lists them, with the exact command to run against each.

6. Run a Function, Endpoint or Workflow

w6w run also takes a fn_…, ep_… or wf_… id, without --action. To address one by its key, prefix the key with fn:, ep: or wf::

w6w run fn:send-welcome-email --payload '{"to":"a@example.com"}'
w6w run ep_01H… --payload '{"email":"a@example.com"}'

A workflow started with w6w run is queued and not waited on. To wait for its result, use w6w workflows run:

w6w workflows run wf_01H… --wait --input '{"email":"a@example.com"}'

You should see the run id, its final status, and its output. A run that’s still going when the server stops waiting is reported as running, which isn’t an error.

Use it in scripts and CI

Add --json to any command to get the raw response, then pipe it to a tool such as jq:

w6w workflows list --json | jq -r '.[] | select(.status == "active") | .id'

Every command exits with a code you can test:

Exit code	Meaning
`0`	Success. Includes a workflow that’s still queued or running.
`1`	A usage or configuration problem: an unknown flag, bad JSON, or no base URL or token.
`2`	The API returned an error, or couldn’t be reached.
`3`	A workflow run finished with status `failed`.

So this step fails your pipeline when the workflow fails:

w6w workflows run wf_01H… --wait --input '{"release":"1.4.2"}'

Errors go to stderr as w6w: <message>. Colour is used only when the output is a terminal. Turn it off with --no-color or by setting NO_COLOR.

Retries

The CLI makes one request per command and never retries. In CI, retry the command yourself only for exit code 2, and only when it’s safe to run twice.

Call a route the CLI doesn’t cover

The CLI covers the operations listed in the command reference. For any other route, use the Node SDK’s or Python client’s request method, which use the same W6W_BASE_URL and W6W_TOKEN.

Known limitations

A few flags shown in --help aren’t applied yet. Use these instead:

Command	What happens	Do this instead
`w6w functions run --id-or-key <key>` or `w6w endpoints run --id-or-key <key>`	The flag isn’t read, so the command says it needs a key.	Pass the key as an argument: `w6w functions run <key>`.
`w6w functions run <key> --inputs '…'`	The inputs are ignored and the Function gets `{}`. `--payload` is rejected.	`w6w run fn:<key> --payload '…'`
`w6w endpoints run <key> --input '…'`	The input is ignored and the Endpoint gets `{}`. `--payload` is rejected.	`w6w run ep:<key> --payload '…'`
`w6w workflows update … --if-unmodified-since …`	The check is skipped, so the save always wins.	Make conditional saves from an SDK with `ifUnmodifiedSince`.

Troubleshooting

Symptom	Cause	What to do
`w6w: No w6w base URL is configured` (exit 1)	`W6W_BASE_URL` is unset or empty.	Export it, or pass `--base-url`.
`w6w: No w6w API token is configured` (exit 1)	`W6W_TOKEN` is unset or empty.	Export it, or pass `--token`.
`w6w: Could not reach the w6w server` (exit 2)	The base URL is wrong or the server is down.	Check the URL, and your network or VPN.
`w6w: unknown flag: --…` (exit 1)	The flag doesn’t exist for that command.	Run the command with `--help` to see its flags.
A `401` (exit 2)	The token is wrong, disabled or revoked.	Create a new token in Studio → Settings → API tokens.
`w6w: No connection found for app` (exit 1)	You have no connection for that app id.	Run `w6w connections list` and check the `APP` column.
`--app` says there are several connections	You have more than one connection for that app.	Use one of the `w6w run conn_…` commands it prints.

Where to next

Command reference: every command and flag.
Node SDK: the same operations from code.