browsix 1.2.0__tar.gz → 1.3.0__tar.gz

{browsix-1.2.0 → browsix-1.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: browsix
-Version: 1.2.0
+Version: 1.3.0
 Summary: Browser automation CLI — wraps cdpwave and bidiwave, no Node.js, no Chromium download
 Project-URL: Homepage, https://github.com/MathiasPaulenko/browsix
 Project-URL: Repository, https://github.com/MathiasPaulenko/browsix
@@ -49,8 +49,9 @@ Description-Content-Type: text/markdown
 [![PyPI](https://img.shields.io/pypi/v/browsix.svg)](https://pypi.org/project/browsix/)
 [![Python](https://img.shields.io/pypi/pyversions/browsix.svg)](https://pypi.org/project/browsix/)
 [![License](https://img.shields.io/github/license/MathiasPaulenko/browsix.svg)](https://github.com/MathiasPaulenko/browsix/blob/main/LICENSE)
+[![Docs](https://img.shields.io/badge/docs-mkdocs-blue.svg)](https://mathiaspaulenko.github.io/browsix/)
 
-> Browser automation CLI — wraps cdpwave and bidiwave. No Node.js, no Chromium download. Uses your existing Chrome/Edge.
+> Browser automation CLI — wraps cdpwave and bidiwave. No Node.js, no Chromium download. Uses your existing Chrome/Edge. 80+ commands across CDP and BiDi backends.
 
 ## Install
 
@@ -64,11 +65,23 @@ pip install browsix[cdp]
 # Take a screenshot
 browsix screenshot https://example.com -o out.png
 
+# Full-page screenshot
+browsix screenshot https://example.com -o full.png --full-page
+
+# Screenshot of a specific element
+browsix screenshot https://example.com -o el.png --selector "h1"
+
 # Generate a PDF
 browsix pdf https://example.com -o out.pdf --paper a4
 
 # Evaluate JavaScript
 browsix eval https://example.com -e "document.title"
+
+# Scrape page content
+browsix scrape https://example.com --selector "article"
+
+# Emulate a device
+browsix device https://example.com --preset iphone-15 -o shot.png
 ```
 
 ## Multi-action
@@ -84,6 +97,9 @@ actions:
   - eval:
       url: https://example.com
       expression: document.title
+  - pdf:
+      url: https://example.com
+      paper: a4
 ```
 
 ```bash
@@ -95,14 +111,61 @@ browsix multi actions.yml
 browsix supports two backends:
 
 - **CDP** (cdpwave) — default, full feature support. `pip install browsix[cdp]`
-- **BiDi** (bidiwave) — minimal, WebDriver BiDi protocol. `pip install browsix[bidi]`
+- **BiDi** (bidiwave) — WebDriver BiDi protocol, 23+ methods implemented. `pip install browsix[bidi]`
 
 Select with `--backend`:
 
 ```bash
-browsix --backend cdp screenshot https://example.com -o out.png
+browsix --backend bidi screenshot https://example.com -o out.png
 ```
 
+### BiDi backend support
+
+The BiDi backend now supports 23+ methods via bidiwave:
+
+| Category | Methods |
+|----------|---------|
+| Navigation | `navigate`, `go_back`, `go_forward`, `reload`, `stop_loading`, `wait_for` |
+| Screenshots | `screenshot`, `screenshot_selector`, `pdf` |
+| Tabs | `list_tabs`, `new_tab`, `close_tab`, `activate_tab` |
+| DOM | `dom_get`, `dom_query`, `dom_set_attr`, `dom_get_attr`, `dom_remove_attr`, `dom_remove`, `dom_focus`, `dom_scroll` |
+| Cookies | `get_cookies`, `set_cookie`, `delete_cookie`, `clear_cookies` |
+| Network | `set_headers`, `set_user_agent`, `block_requests`, `throttle_network`, `set_cache_disabled`, `intercept_requests`, `mock_response` |
+| Emulation | `emulate_device`, `set_viewport`, `set_geolocation`, `set_timezone`, `set_dark_mode`, `set_locale`, `set_touch_emulation`, `set_cpu_throttle`, `set_sensors` |
+| Browser | `browser_version`, `eval`, `raw`, `capture_console`, `capture_logs` |
+| Security | `get_security_state`, `ignore_cert_errors` |
+| Contexts | `new_context`, `list_contexts`, `close_context`, `get_window_bounds`, `set_window_bounds` |
+| Input | `click`, `type_text`, `fill`, `select_option`, `hover`, `key_press`, `drag`, `tap` |
+| Storage | `storage_get`, `storage_set`, `storage_clear`, `storage_list` |
+| Dialogs | `dialog_accept`, `dialog_dismiss`, `grant_permission`, `reset_permissions` |
+
+CDP-only features (HAR, screencast, a11y, downloads, performance profiling,
+CSS inspection, debugging, DOM snapshot, overlay, cache storage, IndexedDB,
+service workers, animations, WebAuthn, WebAudio, Media, Cast, Bluetooth)
+require `--backend cdp`.
+
+## Commands
+
+browsix provides 80+ CLI commands organized into categories:
+
+| Category | Commands |
+|----------|----------|
+| Capture | `screenshot`, `pdf`, `screencast`, `scrape` |
+| Navigate | `navigate`, `back`, `forward`, `reload`, `stop`, `tabs` |
+| Console | `console`, `logs`, `har` |
+| Cookies | `cookies` (get/set/delete/clear) |
+| Network | `headers`, `user-agent`, `block`, `throttle`, `cache`, `intercept`, `mock` |
+| Browser | `open`, `close`, `version` |
+| Emulation | `device`, `viewport`, `geolocation`, `timezone`, `dark-mode` |
+| Input | `click`, `type`, `fill`, `select`, `hover`, `key`, `drag`, `tap` |
+| CSS | `css-styles`, `css-computed`, `css-rules` |
+| Debug | `debug-break`, `debug-step`, `debug-pause`, `debug-resume` |
+| Performance | `perf-metrics`, `perf-trace`, `perf-profile`, `perf-coverage` |
+| Storage | `storage` (get/set/clear/list), `indexeddb` |
+| Advanced | `sw`, `animation`, `record`, `replay`, `webauthn`, `cast`, `bluetooth` |
+
+Run `browsix --help` for the full list.
+
 ## Comparison
 
 | Feature | browsix | shot-scraper | Playwright |
@@ -117,6 +180,10 @@ browsix --backend cdp screenshot https://example.com -o out.png
 | Device emulation | Yes | Yes | Yes |
 | HAR capture | Yes | No | Yes |
 | PDF generation | Yes | Yes | Yes |
+| Network throttling | Yes | No | Yes |
+| Cookie management | Yes | No | Yes |
+| Session recording | Yes | No | No |
+| Serve mode (HTTP API) | Yes | No | No |
 | Shell completions | Yes | No | No |
 
 ## Documentation

browsix-1.3.0/README.md

@@ -0,0 +1,150 @@
+# browsix
+
+[![CI](https://github.com/MathiasPaulenko/browsix/actions/workflows/ci.yml/badge.svg)](https://github.com/MathiasPaulenko/browsix/actions/workflows/ci.yml)
+[![PyPI](https://img.shields.io/pypi/v/browsix.svg)](https://pypi.org/project/browsix/)
+[![Python](https://img.shields.io/pypi/pyversions/browsix.svg)](https://pypi.org/project/browsix/)
+[![License](https://img.shields.io/github/license/MathiasPaulenko/browsix.svg)](https://github.com/MathiasPaulenko/browsix/blob/main/LICENSE)
+[![Docs](https://img.shields.io/badge/docs-mkdocs-blue.svg)](https://mathiaspaulenko.github.io/browsix/)
+
+> Browser automation CLI — wraps cdpwave and bidiwave. No Node.js, no Chromium download. Uses your existing Chrome/Edge. 80+ commands across CDP and BiDi backends.
+
+## Install
+
+```bash
+pip install browsix[cdp]
+```
+
+## Quick start
+
+```bash
+# Take a screenshot
+browsix screenshot https://example.com -o out.png
+
+# Full-page screenshot
+browsix screenshot https://example.com -o full.png --full-page
+
+# Screenshot of a specific element
+browsix screenshot https://example.com -o el.png --selector "h1"
+
+# Generate a PDF
+browsix pdf https://example.com -o out.pdf --paper a4
+
+# Evaluate JavaScript
+browsix eval https://example.com -e "document.title"
+
+# Scrape page content
+browsix scrape https://example.com --selector "article"
+
+# Emulate a device
+browsix device https://example.com --preset iphone-15 -o shot.png
+```
+
+## Multi-action
+
+Create a YAML config and run multiple actions in sequence:
+
+```yaml
+# actions.yml
+actions:
+  - screenshot:
+      url: https://example.com
+      full_page: true
+  - eval:
+      url: https://example.com
+      expression: document.title
+  - pdf:
+      url: https://example.com
+      paper: a4
+```
+
+```bash
+browsix multi actions.yml
+```
+
+## Backends
+
+browsix supports two backends:
+
+- **CDP** (cdpwave) — default, full feature support. `pip install browsix[cdp]`
+- **BiDi** (bidiwave) — WebDriver BiDi protocol, 23+ methods implemented. `pip install browsix[bidi]`
+
+Select with `--backend`:
+
+```bash
+browsix --backend bidi screenshot https://example.com -o out.png
+```
+
+### BiDi backend support
+
+The BiDi backend now supports 23+ methods via bidiwave:
+
+| Category | Methods |
+|----------|---------|
+| Navigation | `navigate`, `go_back`, `go_forward`, `reload`, `stop_loading`, `wait_for` |
+| Screenshots | `screenshot`, `screenshot_selector`, `pdf` |
+| Tabs | `list_tabs`, `new_tab`, `close_tab`, `activate_tab` |
+| DOM | `dom_get`, `dom_query`, `dom_set_attr`, `dom_get_attr`, `dom_remove_attr`, `dom_remove`, `dom_focus`, `dom_scroll` |
+| Cookies | `get_cookies`, `set_cookie`, `delete_cookie`, `clear_cookies` |
+| Network | `set_headers`, `set_user_agent`, `block_requests`, `throttle_network`, `set_cache_disabled`, `intercept_requests`, `mock_response` |
+| Emulation | `emulate_device`, `set_viewport`, `set_geolocation`, `set_timezone`, `set_dark_mode`, `set_locale`, `set_touch_emulation`, `set_cpu_throttle`, `set_sensors` |
+| Browser | `browser_version`, `eval`, `raw`, `capture_console`, `capture_logs` |
+| Security | `get_security_state`, `ignore_cert_errors` |
+| Contexts | `new_context`, `list_contexts`, `close_context`, `get_window_bounds`, `set_window_bounds` |
+| Input | `click`, `type_text`, `fill`, `select_option`, `hover`, `key_press`, `drag`, `tap` |
+| Storage | `storage_get`, `storage_set`, `storage_clear`, `storage_list` |
+| Dialogs | `dialog_accept`, `dialog_dismiss`, `grant_permission`, `reset_permissions` |
+
+CDP-only features (HAR, screencast, a11y, downloads, performance profiling,
+CSS inspection, debugging, DOM snapshot, overlay, cache storage, IndexedDB,
+service workers, animations, WebAuthn, WebAudio, Media, Cast, Bluetooth)
+require `--backend cdp`.
+
+## Commands
+
+browsix provides 80+ CLI commands organized into categories:
+
+| Category | Commands |
+|----------|----------|
+| Capture | `screenshot`, `pdf`, `screencast`, `scrape` |
+| Navigate | `navigate`, `back`, `forward`, `reload`, `stop`, `tabs` |
+| Console | `console`, `logs`, `har` |
+| Cookies | `cookies` (get/set/delete/clear) |
+| Network | `headers`, `user-agent`, `block`, `throttle`, `cache`, `intercept`, `mock` |
+| Browser | `open`, `close`, `version` |
+| Emulation | `device`, `viewport`, `geolocation`, `timezone`, `dark-mode` |
+| Input | `click`, `type`, `fill`, `select`, `hover`, `key`, `drag`, `tap` |
+| CSS | `css-styles`, `css-computed`, `css-rules` |
+| Debug | `debug-break`, `debug-step`, `debug-pause`, `debug-resume` |
+| Performance | `perf-metrics`, `perf-trace`, `perf-profile`, `perf-coverage` |
+| Storage | `storage` (get/set/clear/list), `indexeddb` |
+| Advanced | `sw`, `animation`, `record`, `replay`, `webauthn`, `cast`, `bluetooth` |
+
+Run `browsix --help` for the full list.
+
+## Comparison
+
+| Feature | browsix | shot-scraper | Playwright |
+|---------|---------|--------------|------------|
+| Language | Python | Python | Multi |
+| Node.js required | No | Yes | Yes |
+| Chromium download | No | Yes | Yes |
+| CDP backend | Yes | Yes | Yes |
+| BiDi backend | Yes | No | No |
+| CLI-first | Yes | Yes | No |
+| Multi-action YAML | Yes | No | No |
+| Device emulation | Yes | Yes | Yes |
+| HAR capture | Yes | No | Yes |
+| PDF generation | Yes | Yes | Yes |
+| Network throttling | Yes | No | Yes |
+| Cookie management | Yes | No | Yes |
+| Session recording | Yes | No | No |
+| Serve mode (HTTP API) | Yes | No | No |
+| Shell completions | Yes | No | No |
+
+## Documentation
+
+Full docs at [mathiaspaulenko.github.io/browsix](https://mathiaspaulenko.github.io/browsix/).
+
+## License
+
+MIT

{browsix-1.2.0 → browsix-1.3.0}/browsix/backend/bidi.py

@@ -38,18 +38,21 @@ class BiDiBackend(AbstractBackend):
     Supports: launch, navigate, screenshot, screenshot_selector, pdf, eval,
     raw, close, go_back, go_forward, reload, stop_loading, wait_for,
     list_tabs, new_tab, close_tab, activate_tab, capture_console,
-    capture_logs, DOM methods, storage methods, cookies (get/set/delete/clear),
-    set_headers, set_user_agent, browser_version, emulate_device, set_viewport,
-    set_geolocation, set_timezone, set_dark_mode, set_locale, set_touch_emulation,
+    capture_logs, DOM methods, dom_snapshot, storage methods,
+    cookies (get/set/delete/clear), set_headers, set_user_agent,
+    browser_version, emulate_device, set_viewport, set_geolocation,
+    set_timezone, set_dark_mode, set_locale, set_touch_emulation,
     set_cpu_throttle, set_sensors, new_context, list_contexts, close_context,
     get_window_bounds, set_window_bounds, get_security_state, ignore_cert_errors,
+    perf_metrics, perf_coverage, css_get_styles, css_get_computed,
     dialog_accept, dialog_dismiss, grant_permission, reset_permissions,
     click, type_text, fill, select_option, hover, key_press, drag, tap,
     block_requests, throttle_network, set_cache_disabled, intercept_requests,
     mock_response.
 
-    CDP-only features (HAR, screencast, a11y, downloads, performance,
-    CSS, debug, DOM snapshot, overlay, cache storage, IndexedDB,
+    CDP-only features (HAR, screencast, a11y, downloads, perf_trace,
+    perf_profile, perf_heap_snapshot, perf_css_coverage, css_get_stylesheets,
+    css_get_rules, debug, overlay, cache storage, IndexedDB,
     service workers, animations, WebAuthn, WebAudio, Media, Cast, Bluetooth)
     raise NotImplementedError.
     """
@@ -1013,7 +1016,37 @@ class BiDiBackend(AbstractBackend):
     # ── Performance ───────────────────────────────────────
 
     async def perf_metrics(self) -> dict[str, Any]:
-        raise NotImplementedError("perf_metrics is not supported by BiDiBackend")
+        """Get performance metrics via JS Performance API.
+
+        Uses script.evaluate to collect performance.timing,
+        performance.memory, and navigation timing data.
+
+        Returns:
+            Dict mapping metric names to values.
+        """
+        if self._client is None or self._context is None:
+            raise RuntimeError("BiDiBackend not launched. Call launch() first.")
+        js = (
+            "JSON.stringify({"
+            "  navigationStart: performance.timing.navigationStart,"
+            "  loadEventEnd: performance.timing.loadEventEnd,"
+            "  domContentLoadedEventEnd: performance.timing.domContentLoadedEventEnd,"
+            "  responseEnd: performance.timing.responseEnd,"
+            "  domInteractive: performance.timing.domInteractive,"
+            "  domComplete: performance.timing.domComplete,"
+            "  loadEventStart: performance.timing.loadEventStart,"
+            "  jsHeapUsedSize: performance.memory ? performance.memory.usedJSHeapSize : null,"
+            "  jsHeapTotalSize: performance.memory ? performance.memory.totalJSHeapSize : null,"
+            "  jsHeapLimit: performance.memory ? performance.memory.jsHeapSizeLimit : null,"
+            "  resourceCount: performance.getEntriesByType('resource').length,"
+            "  transferSize: performance.getEntriesByType('resource')"
+            "    .reduce((s,e)=>s+(e.transferSize||0),0)"
+            "})"
+        )
+        result = await self._client.script.evaluate(self._context, js)
+        import json as _json
+        val = result.value if hasattr(result, "value") else result
+        return _json.loads(val) if isinstance(val, str) else dict(val)
 
     async def perf_trace(self, duration_ms: int = 3000) -> dict[str, Any]:
         raise NotImplementedError("perf_trace is not supported by BiDiBackend")
@@ -1027,7 +1060,22 @@ class BiDiBackend(AbstractBackend):
         )
 
     async def perf_coverage(self) -> dict[str, Any]:
-        raise NotImplementedError("perf_coverage is not supported by BiDiBackend")
+        """Get JS coverage data via CDP Profiler.
+
+        Uses CDP bridge to enable profiler and take precise coverage.
+
+        Returns:
+            Dict with 'result' key containing script coverage entries.
+        """
+        if self._client is None:
+            raise RuntimeError("BiDiBackend not launched. Call launch() first.")
+        await self._client.cdp.send_command("Profiler.enable", {})
+        await self._client.cdp.send_command(
+            "Profiler.startPreciseCoverage",
+            {"callCount": True, "detailed": True},
+        )
+        result = await self._client.cdp.send_command("Profiler.takePreciseCoverage", {})
+        return dict(result) if result else {}
 
     async def perf_css_coverage(self) -> dict[str, Any]:
         raise NotImplementedError(
@@ -1037,7 +1085,49 @@ class BiDiBackend(AbstractBackend):
     # ── CSS ────────────────────────────────────────────────
 
     async def css_get_styles(self, selector: str) -> dict[str, Any]:
-        raise NotImplementedError("css_get_styles is not supported by BiDiBackend")
+        """Get inline and matched styles for an element via JS.
+
+        Uses script.evaluate to extract inline styles and matched
+        CSS rules from document.styleSheets for the given selector.
+
+        Args:
+            selector: CSS selector for the target element.
+
+        Returns:
+            Dict containing inlineStyles and matchedStyles.
+        """
+        if self._client is None or self._context is None:
+            raise RuntimeError("BiDiBackend not launched. Call launch() first.")
+        escaped = selector.replace("'", "\\'")
+        js = (
+            f"(function(){{"
+            f"  var el=document.querySelector('{escaped}');"
+            f"  if(!el) return null;"
+            f"  var inline=el.getAttribute('style')||'';"
+            f"  var matched=[];"
+            f"  for(var i=0;i<document.styleSheets.length;i++){{"
+            f"    try{{"
+            f"      var sheet=document.styleSheets[i];"
+            f"      var rules=sheet.cssRules||sheet.rules;"
+            f"      for(var j=0;j<rules.length;j++){{"
+            f"        if(el.matches(rules[j].selectorText)){{"
+            f"          matched.push({{"
+            f"            selectorText:rules[j].selectorText,"
+            f"            cssText:rules[j].cssText"
+            f"          }});"
+            f"        }}"
+            f"      }}"
+            f"    }}catch(e){{}}"
+            f"  }}"
+            f"  return JSON.stringify({{inlineStyles:inline,matchedStyles:matched}});"
+            f"}})()"
+        )
+        result = await self._client.script.evaluate(self._context, js)
+        val = result.value if hasattr(result, "value") else result
+        if not val:
+            raise RuntimeError(f"Element not found: {selector}")
+        import json as _json
+        return _json.loads(val) if isinstance(val, str) else dict(val)
 
     async def css_get_stylesheets(self) -> list[dict[str, Any]]:
         raise NotImplementedError(
@@ -1048,9 +1138,36 @@ class BiDiBackend(AbstractBackend):
         raise NotImplementedError("css_get_rules is not supported by BiDiBackend")
 
     async def css_get_computed(self, selector: str) -> dict[str, Any]:
-        raise NotImplementedError(
-            "css_get_computed is not supported by BiDiBackend"
+        """Get computed styles for an element via JS getComputedStyle.
+
+        Args:
+            selector: CSS selector for the target element.
+
+        Returns:
+            Dict mapping CSS property names to computed values.
+        """
+        if self._client is None or self._context is None:
+            raise RuntimeError("BiDiBackend not launched. Call launch() first.")
+        escaped = selector.replace("'", "\\'")
+        js = (
+            f"(function(){{"
+            f"  var el=document.querySelector('{escaped}');"
+            f"  if(!el) return null;"
+            f"  var cs=getComputedStyle(el);"
+            f"  var result={{}};"
+            f"  for(var i=0;i<cs.length;i++){{"
+            f"    var prop=cs.item(i);"
+            f"    result[prop]=cs.getPropertyValue(prop);"
+            f"  }}"
+            f"  return JSON.stringify(result);"
+            f"}})()"
         )
+        result = await self._client.script.evaluate(self._context, js)
+        val = result.value if hasattr(result, "value") else result
+        if not val:
+            raise RuntimeError(f"Element not found: {selector}")
+        import json as _json
+        return _json.loads(val) if isinstance(val, str) else dict(val)
 
     # ── Debugging ──────────────────────────────────────────
 
@@ -1094,7 +1211,19 @@ class BiDiBackend(AbstractBackend):
     # ── DOM Snapshot ───────────────────────────────────────
 
     async def dom_snapshot(self) -> dict[str, Any]:
-        raise NotImplementedError("dom_snapshot is not supported by BiDiBackend")
+        """Capture a DOM snapshot via JS.
+
+        Uses script.evaluate to serialize the full DOM tree.
+
+        Returns:
+            Dict containing 'html' with the full outerHTML.
+        """
+        if self._client is None or self._context is None:
+            raise RuntimeError("BiDiBackend not launched. Call launch() first.")
+        js = "document.documentElement.outerHTML"
+        result = await self._client.script.evaluate(self._context, js)
+        html = result.value if hasattr(result, "value") else result
+        return {"html": str(html)}
 
     # ── Overlay ────────────────────────────────────────────
 

{browsix-1.2.0 → browsix-1.3.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "browsix"
-version = "1.2.0"
+version = "1.3.0"
 description = "Browser automation CLI — wraps cdpwave and bidiwave, no Node.js, no Chromium download"
 readme = "README.md"
 license = { text = "MIT" }

{browsix-1.2.0 → browsix-1.3.0}/tests/unit/test_bidi_backend.py

@@ -95,6 +95,16 @@ class TestBiDiBackend:
                 await backend.set_cpu_throttle(4.0)
             with pytest.raises(RuntimeError, match="not launched"):
                 await backend.set_sensors(MagicMock())
+            with pytest.raises(RuntimeError, match="not launched"):
+                await backend.perf_metrics()
+            with pytest.raises(RuntimeError, match="not launched"):
+                await backend.perf_coverage()
+            with pytest.raises(RuntimeError, match="not launched"):
+                await backend.css_get_styles("h1")
+            with pytest.raises(RuntimeError, match="not launched"):
+                await backend.css_get_computed("h1")
+            with pytest.raises(RuntimeError, match="not launched"):
+                await backend.dom_snapshot()
 
     async def test_bidi_paridad_methods_raise_runtime_without_launch(self) -> None:
         with patch("browsix.backend.bidi.BiDiClient", MagicMock()):

browsix-1.2.0/README.md

@@ -1,83 +0,0 @@
-# browsix
-
-[![CI](https://github.com/MathiasPaulenko/browsix/actions/workflows/ci.yml/badge.svg)](https://github.com/MathiasPaulenko/browsix/actions/workflows/ci.yml)
-[![PyPI](https://img.shields.io/pypi/v/browsix.svg)](https://pypi.org/project/browsix/)
-[![Python](https://img.shields.io/pypi/pyversions/browsix.svg)](https://pypi.org/project/browsix/)
-[![License](https://img.shields.io/github/license/MathiasPaulenko/browsix.svg)](https://github.com/MathiasPaulenko/browsix/blob/main/LICENSE)
-
-> Browser automation CLI — wraps cdpwave and bidiwave. No Node.js, no Chromium download. Uses your existing Chrome/Edge.
-
-## Install
-
-```bash
-pip install browsix[cdp]
-```
-
-## Quick start
-
-```bash
-# Take a screenshot
-browsix screenshot https://example.com -o out.png
-
-# Generate a PDF
-browsix pdf https://example.com -o out.pdf --paper a4
-
-# Evaluate JavaScript
-browsix eval https://example.com -e "document.title"
-```
-
-## Multi-action
-
-Create a YAML config and run multiple actions in sequence:
-
-```yaml
-# actions.yml
-actions:
-  - screenshot:
-      url: https://example.com
-      full_page: true
-  - eval:
-      url: https://example.com
-      expression: document.title
-```
-
-```bash
-browsix multi actions.yml
-```
-
-## Backends
-
-browsix supports two backends:
-
-- **CDP** (cdpwave) — default, full feature support. `pip install browsix[cdp]`
-- **BiDi** (bidiwave) — minimal, WebDriver BiDi protocol. `pip install browsix[bidi]`
-
-Select with `--backend`:
-
-```bash
-browsix --backend cdp screenshot https://example.com -o out.png
-```
-
-## Comparison
-
-| Feature | browsix | shot-scraper | Playwright |
-|---------|---------|--------------|------------|
-| Language | Python | Python | Multi |
-| Node.js required | No | Yes | Yes |
-| Chromium download | No | Yes | Yes |
-| CDP backend | Yes | Yes | Yes |
-| BiDi backend | Yes | No | No |
-| CLI-first | Yes | Yes | No |
-| Multi-action YAML | Yes | No | No |
-| Device emulation | Yes | Yes | Yes |
-| HAR capture | Yes | No | Yes |
-| PDF generation | Yes | Yes | Yes |
-| Shell completions | Yes | No | No |
-
-## Documentation
-
-Full docs at [mathiaspaulenko.github.io/browsix](https://mathiaspaulenko.github.io/browsix/).
-
-## License
-
-MIT