scraperapi-cli

ScraperAPI CLI (sapi)

sapi is the official ScraperAPI command-line tool. It is the right choice when:

• The user is already in a terminal and wants a result now without writing a script.
• A scrape is part of a shell pipeline (sapi … | jq …, xargs, make, GitHub Actions).
• A one-off scheduled task (cron, launchd, systemd timer) needs to hit ScraperAPI without a project setup.
• The user is exploring — testing whether --render or --premium unblocks a target before committing the choice to code.

If the user is writing application code in Python, Node, PHP, Ruby, or Java, point them at the matching SDK skill instead — the CLI is for shells, not application logic.

Install and authenticate

npm install -g scraperapi-cli    # requires Node.js 18+
sapi init                        # interactive: prompts for the key and validates it

Non-interactive setup (for CI / Dockerfiles):

sapi init --api-key "$SCRAPERAPI_API_KEY"

Key resolution order

sapi looks for the API key in this order, stopping at the first hit:

1. --api-key <key> flag on the command
2. SCRAPERAPI_API_KEY environment variable
3. ~/.config/scraperapi/config.json (written by sapi init)

In CI, prefer the env var — it keeps the key out of shell history and config files.

Output contract — important for piping

Stream	What goes there
stdout	The data (page body, JSON, table rows)
stderr	Spinners, warnings, errors

When stdout is not a TTY (a pipe or redirect), sapi automatically switches to JSON mode:

sapi scrape https://example.com | jq .body            # auto-JSON
sapi scrape https://example.com > out.json            # auto-JSON
sapi scrape https://example.com                       # human output (page body to stdout)

Force JSON mode in a TTY with --json. Force a specific body format with --output html|markdown|text|json|csv — --output overrides the non-TTY auto-JSON rule, so you can pipe raw HTML or markdown into another tool without it being wrapped in JSON.