Reader CLI

Read the web from your terminal. Built for AI coding agents.

The Reader CLI scrapes, crawls, and screenshots web pages via the Reader API. It outputs clean markdown to stdout, ready to pipe into any AI workflow.

Install

npm install -g @vakra-dev/reader-cli

Or run directly:

npx @vakra-dev/reader-cli scrape https://example.com

Setup

For humans - save your API key once:

reader config set api-key rdr_your_key_here

For AI agents / CI - set the environment variable:

export READER_API_KEY=rdr_your_key_here

Get your API key at console.reader.dev.

Usage

Scrape a page

reader scrape https://example.com

Output is clean markdown, printed to stdout:

reader scrape https://docs.stripe.com/payments > stripe-payments.md

Formats

reader scrape https://example.com                     # markdown (default)
reader scrape https://example.com -f html              # cleaned HTML
reader scrape https://example.com -f screenshot -o page.png   # full-page screenshot

JSON output

Get the full API response with metadata:

reader scrape https://example.com --json

Crawl a site

Discover and scrape all pages:

reader crawl https://docs.example.com --max-pages 20

Save pages to individual files:

reader crawl https://docs.example.com -o ./docs/

List URLs only:

reader crawl https://example.com --urls-only

Check status

reader status

Reader CLI v0.1.0
API:     https://api.reader.dev (connected)
Key:     rdr_...c3bb
Credits: 874 / 1000 (free tier)
Resets:  2026-07-01

Check credits

reader credits
reader credits --json

Configuration

reader config set api-key <key>     # save API key
reader config set api-url <url>     # custom API URL
reader config show                  # show current config

Config is stored at ~/.reader/config.json. Environment variables take precedence:

Setting	Env var	Default
API key	READER_API_KEY	-
API URL	READER_API_URL	https://api.reader.dev

Command reference

reader scrape <url>

Flag	Description	Default
-f, --format	Output format: markdown, html, screenshot	markdown
--json	Full JSON response	-
-o, --output	Write to file	stdout
--no-main-content	Include nav, header, footer	-
--include-tags	CSS selectors to keep	-
--exclude-tags	CSS selectors to remove	-
--wait-for	Wait for CSS selector	-
--timeout	Timeout in ms	30000
--proxy-mode	standard, stealth, auto	auto

reader crawl <url>

Flag	Description	Default
--max-depth	Crawl depth	2
--max-pages	Max pages to crawl	20
--urls-only	Only list URLs	-
--json	Full JSON response	-
-o, --output-dir	Write pages to directory	stdout

reader status

Shows CLI version, API connectivity, credit balance, and tier.

reader credits

Flag	Description
--json	Full JSON response

reader config

Subcommand	Description
set api-key <key>	Save API key
set api-url <url>	Set custom API URL
show	Display current config

For AI agents

The CLI is designed for seamless use with AI coding agents like Claude Code, Cursor, and Codex:

• Markdown to stdout - agents read content directly
• Errors to stderr - never pollutes piped output
• Exit codes - 0 on success, 1 on error
• No interactive prompts - everything via flags and env vars
• --json flag - structured output for programmatic use

License

MIT - see LICENSE.