Obscura Share Help Try It

Lightweight headless browser engine written in Rust for web scraping and AI agent automation, with a Chrome DevTools Protocol server compatible with Puppeteer and Playwright.

Category

Network

License

Apache 2

Platforms

Windows macOS Linux

Released

2026

Path

c:\obscura\obscura.exe

Benefits

Single static binary with no Chrome or Node.js dependency. Uses ~30 MB of memory and loads typical pages in under 100 ms (roughly 5-10x lighter and faster than headless Chrome). Implements the Chrome DevTools Protocol so existing Puppeteer and Playwright scripts can connect unchanged. Optional stealth mode randomizes browser fingerprints (GPU, canvas, audio, userAgentData) and blocks 3,520+ tracker domains.

Notes

Three subcommands: 'fetch' renders a single URL, 'serve' starts a CDP WebSocket server (default port 9222), and 'scrape' runs many URLs in parallel. There is no --version flag in v0.1.1; check the binary against the GitHub releases page. Stealth and tracker-blocking features require a build compiled with --features stealth. Only public, internet-reachable hostnames can be scanned; respect robots.txt with --obey-robots and the target site's terms of use.

Used by

Web scraping pipelines, AI agent automation, drop-in replacement for headless Chrome with Puppeteer/Playwright

Try it

• Fetch a page and extract its title
  Render https://example.com and evaluate 'document.title' against the loaded DOM. example.com is reserved by IANA (RFC 2606) for documentation use, so it is safe to hit from demos.
• Fetch a page and dump rendered text
  Render https://example.com and dump the visible text content (no HTML tags). Useful for feeding scraped content into text-processing pipelines.
• Fetch a page and extract all links
  Render https://example.com and dump every anchor href on the page. The 'links' dump mode is the easiest way to enumerate outbound links without parsing HTML yourself.

Agree to terms to run demos.

# Obscura - Help

Source: https://github.com/h4ckf0r0day/obscura

## Top-level

```
Obscura - A lightweight headless browser for web scraping and automation

Usage: obscura.exe [OPTIONS] [COMMAND]

Commands:
  serve
  fetch
  scrape
  help    Print this message or the help of the given subcommand(s)

Options:
  -v, --verbose
  -p, --port <PORT>              [default: 9222]
      --proxy <PROXY>
      --obey-robots
      --user-agent <USER_AGENT>
  -h, --help                     Print help
```

## obscura fetch

```
Usage: obscura.exe fetch [OPTIONS] <URL>

Arguments:
  <URL>

Options:
      --dump <DUMP>              [default: html] [possible values: html, text, links]
  -v, --verbose
      --selector <SELECTOR>
      --wait <WAIT>              [default: 5]
      --wait-until <WAIT_UNTIL>  [default: load]
      --user-agent <USER_AGENT>
      --stealth
  -e, --eval <EVAL>
  -q, --quiet
  -h, --help                     Print help
```

## obscura serve

```
Usage: obscura.exe serve [OPTIONS]

Options:
  -p, --port <PORT>              [default: 9222]
  -v, --verbose
      --proxy <PROXY>
      --user-agent <USER_AGENT>
      --stealth
      --workers <WORKERS>        [default: 1]
  -h, --help                     Print help
```

## obscura scrape

```
Usage: obscura.exe scrape [OPTIONS] [URLS]...

Arguments:
  [URLS]...

Options:
  -e, --eval <EVAL>
  -v, --verbose
      --concurrency <CONCURRENCY>  [default: 10]
      --format <FORMAT>            [default: json]
      --timeout <TIMEOUT>          [default: 60]
  -h, --help                       Print help
```

## Typical invocations

Fetch a page and dump rendered HTML:

```
obscura fetch https://example.com
```

Get the page title via JavaScript evaluation:

```
obscura fetch https://example.com --eval "document.title" --quiet
```

Extract all links:

```
obscura fetch https://example.com --dump links --quiet
```

Wait for dynamic content to settle before dumping:

```
obscura fetch https://example.com --wait-until networkidle0 --dump html
```

Start a CDP WebSocket server on the default port for Puppeteer/Playwright:

```
obscura serve --port 9222 --stealth
```

Scrape multiple URLs in parallel and emit JSON:

```
obscura scrape https://example.com https://example.org --concurrency 5 --eval "document.title" --format json
```

## Notes

- v0.1.1 has no `--version` or `version` subcommand. Compare the binary's
  modification date against the GitHub releases page to identify the version.
- Stealth mode (`--stealth`) and tracker blocking require a build compiled with
  `--features stealth`.
- `serve` is a long-running daemon. Do not invoke it from a synchronous CFML
  request unless you intend the request to block until the server exits.
- Only public, internet-reachable hostnames can be reached; respect robots.txt
  with `--obey-robots` and the target site's terms of use.