CLI

The Sila CLI is the @sila-sh/cli package. It needs Node.js 20 or later and has no other dependencies.

npx -y @sila-sh/cli@latest --help

Install it globally to get the sila command:

npm install -g @sila-sh/cli

Get set up

sila init logs you in, installs the Sila agent skills, and prints the MCP setup lines.

npx -y @sila-sh/cli@latest init --all --browser

Flag	Description
`--all`	Install the skills for every supported agent without asking.
`--agent <name>`	Install the skills for one agent. Repeat the flag for more agents.
`--browser`	Open the login page in your browser.
`--skip-auth`	Skip login.
`--skip-skills`	Skip the skills install.

init skips login when you already have a saved key that works. Without --all or --agent, the skills installer asks which agents to use.

Log in

sila login

The CLI prints a link, or opens it with --browser. Open the link, check the terminal name on the page, and click Authorize. The CLI waits up to 10 minutes, then saves a new API key. To use an API key you already have:

sila login --api-key sila_your_key

The CLI checks the key before it saves it.

Command	Description
`sila login [--browser] [--api-key <key>]`	Log in with the browser or with a key.
`sila logout`	Remove the saved key. The key stays valid until you revoke it on the API keys page.
`sila status`	Show the CLI version, where the key comes from, your plan, and credits used and remaining. `sila --status` does the same.

Scrape a page

sila scrape https://example.com

Flag	Description
`-f <formats>`	Comma-separated list of `markdown`, `html`, `signals`, `siteInfo`, `screenshot`. Default `markdown`.
`--render`	`auto`, `never`, or `always`. Default `auto`.
`-o <file>`	Write the output to a file instead of the terminal.
`--json`	Print the full JSON response.

With a single markdown or html format, the CLI prints the raw text. With any other formats, or with --json, it prints formatted JSON.

sila scrape https://example.com -f markdown,signals
sila scrape https://example.com -f html -o page.html

See Scrape a page for the response and Credits and plans for costs.

Crawl a site

sila crawl https://example.com --limit 50 --wait -o ./out

Flag	Description
`--limit <n>`	Maximum pages to crawl.
`--no-robots`	Ignore `robots.txt`.
`--wait`	Wait for the crawl to finish, then fetch every page.
`-o <dir>`	Write each page as a markdown file under `<dir>/<host>/`. The home page is `index.md`. Implies `--wait`.

Without --wait or -o, the command prints the crawl ID and returns. With --wait and no -o, it prints a summary instead of writing files. Manage a crawl by ID:

Command	Description
`sila crawl status <id>`	Show the crawl status and stats.
`sila crawl pages <id> [--cursor <n>]`	Show one batch of stored pages. Pass `nextCursor` as `--cursor` for the next batch.
`sila crawl cancel <id>`	Stop a queued or running crawl.

See Create a crawl for how crawls work.

Where the CLI looks for your key

The CLI uses the first key it finds:

The --api-key flag.
The SILA_API_KEY environment variable.
The saved login in ~/.config/sila/config.json. If XDG_CONFIG_HOME is set, the file is $XDG_CONFIG_HOME/sila/config.json.

sila login writes the file with owner-only permissions.

Variable	Description
`SILA_API_KEY`	API key to use.
`SILA_API_URL`	API base URL. Default `https://api.sila.sh`.
`SILA_APP_URL`	Dashboard URL used for login. Default `https://sila.sh`.

Exit codes

Code	Meaning
`0`	Success.
`1`	The command failed. The CLI prints the API error as `code: message`. See Errors.
`2`	Authentication failed. Run `sila login`.

Getting started

AI agents

API reference

Guides

Get set up

Log in

Scrape a page

Crawl a site

Where the CLI looks for your key

Exit codes

​Get set up

​Log in

​Scrape a page

​Crawl a site

​Where the CLI looks for your key

​Exit codes

Get set up

Log in

Scrape a page

Crawl a site

Where the CLI looks for your key

Exit codes