How to Download a Website’s HTML Code
Save a page's original HTML with curl, inspect rendered DOM in browser tools, or mirror its assets for offline use.
There are three different jobs people mean by “download a website’s HTML code”:
- Save the original HTML response: use
curlor another HTTP client. - Inspect the page in a browser: use View Source for the server response or developer tools for the rendered DOM.
- Keep a page usable offline: mirror the HTML and its linked assets instead of saving one file.
For one page, the shortest reliable command is:
curl -o page.html https://example.com/The -o option writes the response to the filename you choose. This saves the HTML response; it does not automatically download every stylesheet, image, font, script, or linked page.
1. Save the original HTML with curl
curl’s official tutorial documents saving a webpage under an explicit filename with -o.
Basic command
curl -o page.html https://example.com/Replace the URL and filename:
curl -o product.html https://shop.example/products/widget
curl -o article.html 'https://example.com/article?id=42'Quote URLs containing &, spaces, or shell metacharacters.
Use the remote filename
Use uppercase -O when you want curl to use the filename from the URL:
curl -O https://example.com/index.htmlIf the URL ends in a slash or does not provide a useful filename, choose one explicitly with -o.
Follow redirects
Many sites redirect HTTP to HTTPS or redirect old paths. Add -L:
curl -L -o page.html https://example.com/old-pathWithout -L, curl may save a small redirect response instead of the destination page.
See headers and status
curl -L -D headers.txt -o page.html https://example.com/
cat headers.txtFor a compact diagnostic without writing the body:
curl -L -I https://example.com/Check the final status, Content-Type, redirects, compression, and caching headers. A successful TCP connection does not guarantee that the response is HTML.
Preserve compressed responses correctly
Ask the server for compression and let curl decompress it before writing the file:
curl -L --compressed -o page.html https://example.com/Set a timeout and fail on HTTP errors
curl --fail --show-error --location \
--connect-timeout 10 --max-time 60 \
-o page.html https://example.com/--fail makes HTTP 4xx and 5xx responses errors, --show-error keeps the diagnostic, and the timeout flags prevent a stalled request from running forever.
Send a browser-like user agent when required
curl -L \
-A 'Mozilla/5.0 (compatible; HTML downloader)' \
-o page.html https://example.com/A user agent can change what a server returns, but it does not bypass authentication, bot checks, or access controls. Respect the site’s terms and robots policy.
2. Download HTML with Python
Python’s requests library saves the response body as text. This example follows redirects, checks the status, and writes UTF-8 text.
import requests
url = "https://example.com/"
response = requests.get(
url,
headers={"User-Agent": "HTML downloader"},
timeout=60,
)
response.raise_for_status()
# response.text decodes using the server's declared encoding.
with open("page.html", "w", encoding=response.encoding or "utf-8") as f:
f.write(response.text)
print(response.status_code, response.url)
Install the dependency with python -m pip install requests. For a binary-safe copy of the exact body, write bytes instead:
with open("page.html", "wb") as f:
f.write(response.content)
Use response.text when you need decoded text; use response.content when you want the raw response bytes.
3. Download HTML with Node.js
Modern Node.js versions include fetch. This complete script follows the server’s response but does not implement browser JavaScript execution.
import { writeFile } from "node:fs/promises";
const url = "https://example.com/";
const controller = new AbortController();
const timer = setTimeout(() => controller.abort(), 60_000);
try {
const response = await fetch(url, {
redirect: "follow",
headers: { "user-agent": "HTML downloader" },
signal: controller.signal,
});
if (!response.ok) {
throw new Error(`HTTP ${response.status} ${response.statusText}`);
}
await writeFile("page.html", await response.text(), "utf8");
console.log(`Saved ${response.url}`);
} finally {
clearTimeout(timer);
}
Save it as download.mjs and run node download.mjs. The fetch API returns the server response; it does not render a JavaScript application.
4. Inspect HTML in a browser
Downloading a response and inspecting a browser page answer different questions.
| Need | Use | What you see |
|---|---|---|
| Original server response | View Source / show-source | HTML initially returned before scripts and other resources run |
| Final page structure | Developer tools → Elements | The DOM after scripts, framework rendering, and browser changes |
Google documents this distinction in its guidance on viewing source and rendered HTML. A JavaScript-heavy page can have very little useful markup in View Source while containing a complete interface in the Elements panel.
View Source
- Open the page in your browser.
- Choose View Source, or enter
view-source:https://example.com/in the address bar. - Save the source with the browser’s save command if you need a local copy.
Inspect the rendered DOM
- Open developer tools.
- Select the Elements or Inspector panel.
- Right-click the root element and choose the copy or save option available in your browser.
The rendered DOM is not necessarily a faithful downloadable source file: browser extensions, user state, JavaScript, consent choices, and runtime data can all affect it.
5. Download a page for offline viewing
One HTML file usually contains links to external resources. To create an offline copy, use a mirroring workflow that rewrites links and downloads dependencies. A typical curl request is insufficient for this job.
Before mirroring, decide whether you need:
- Only the initial HTML response.
- The page plus images, CSS, fonts, and scripts.
- Several pages or an entire site.
- Authenticated or personalized content.
Mirrors can become large, miss resources loaded by JavaScript, or copy private data. Follow the site’s terms, rate limits, and access rules. Test the result by disconnecting from the network and opening the local entry file.
6. What a downloaded HTML file contains
The file can include elements such as headings, links, forms, inline styles, inline scripts, and references to external resources:
<link rel="stylesheet" href="styles.css">
<img src="images/hero.jpg" alt="...">
<script src="app.js" defer></script>
Those references are not the referenced files themselves. Relative URLs may also stop working when the file is opened from file://. An HTML response can therefore be valid while its offline copy looks unstyled or incomplete.
7. JavaScript-rendered pages and dynamic content
HTTP clients such as curl, Requests, and Node fetch do not run page JavaScript. They receive the server response. If the site renders content in the browser, use developer tools to inspect the final DOM or use a browser automation tool that waits for the application to finish rendering.
Common signs that you downloaded only a shell include a nearly empty root element, script bundles with no visible content, or data that appears only after an API call. Save the initial HTML and inspect the browser’s Network panel to identify the requests that populate the page.
8. Troubleshooting
| Symptom | Likely cause | Fix |
|---|---|---|
| Saved file is a redirect or error page | Redirects were not followed or the server returned an error | Use -L, inspect headers, and use --fail |
| File is empty or truncated | Timeout, interrupted transfer, or server connection failure | Set --max-time, retry carefully, and check the exit code |
| Garbled characters | Incorrect character decoding | Inspect the response charset; write raw bytes or use the declared encoding |
| HTML has no visible application | Content is rendered by JavaScript | Use View Source for the initial response and Elements or browser automation for the rendered DOM |
| Images and styles are missing offline | Only the HTML document was downloaded | Use a resource-aware mirroring workflow and verify relative paths |
| 403, 429, or CAPTCHA | Access policy, rate limiting, or bot protection | Authenticate where permitted, slow requests, or use the site’s approved export/API |
| Different content than the browser | Cookies, headers, location, device, or personalization differ | Compare request headers and cookies; do not assume curl reproduces a browser session |
9. Reliability, performance, and cost
- Reliability: record the final URL, status code, timestamp, and response headers alongside the file when reproducibility matters.
- Performance: reuse connections for batches, set finite timeouts, and avoid downloading assets you do not need.
- Retries: retry transient network failures with backoff; avoid rapidly retrying 401, 403, 404, or 429 responses.
- Integrity: calculate a checksum if the file is an input to another build or archival process.
- Cost: direct HTTP downloads consume your network and server resources. Full mirrors can multiply transfer, storage, and request counts.
curl -L --fail --show-error --retry 3 --retry-delay 2 \
--connect-timeout 10 --max-time 60 \
-o page.html https://example.com/Retries are most useful for temporary connection failures. Keep concurrency low enough to respect the site and your own bandwidth.
10. Or skip the browser setup
If your real goal is a clean visual capture rather than the HTML source, ScreenshotNeo returns a screenshot or PDF from one GET request. See the ScreenshotNeo API documentation for all options.
curl -G "https://api.screenshotneo.com/v1/shot" -d access_key=YOUR_API_KEY --data-urlencode url=https://stripe.com -o shot.webpimport requests
r = requests.get("https://api.screenshotneo.com/v1/shot", params={"access_key": "YOUR_API_KEY", "url": "https://stripe.com"}, timeout=90)
open("shot.webp", "wb").write(r.content)const q = new URLSearchParams({ access_key: 'YOUR_API_KEY', url: 'https://stripe.com' }); const res = await fetch(`https://api.screenshotneo.com/v1/shot?${q}`);Before capture, ScreenshotNeo accepts cookie and consent banners and removes more than 60 known consent platforms, newsletter popups, and chat widgets; each step can be disabled. Bot checks, blank pages, timeouts, failed loads, and cache hits are not billed, and response headers identify the page verdict and billing result. Its MCP server lets Claude, Cursor, and other MCP clients use take_screenshot, get_page_info, and capture_pdf. The Free plan includes 1,000 screenshots per month with no card; paid plans start at $5 for 3,000 shots.
Create a free ScreenshotNeo account to try it without a card.
11. FAQ
Does downloading HTML download the whole website?
No. It downloads one HTTP response. Linked resources and other pages require additional requests or a mirroring tool.
Should I use View Source or Elements?
Use View Source for the original response and Elements for the DOM after JavaScript and browser processing.
Why does curl show less content than my browser?
The browser runs JavaScript and sends cookies and other context. curl normally does none of those things.
Can I download a page that requires login?
Only with authorized credentials or an approved export/API. A public URL alone does not grant access to private content.
What filename should I use?
Use a descriptive name such as page.html with -o; use -O only when the URL’s remote filename is suitable.
