Wayback Machine 429 Too Many Requests: Why It Happens and How to Fix It

A Wayback Machine 429 error means the Internet Archive has decided that your browser, script or IP address sent too many requests in a short time, so it has stopped answering for a while. The message is usually “429 Too Many Requests”. It is not a sign that the archive has lost the page, and it is rarely permanent. In most cases you fix it by stopping, waiting, and then sending requests far more slowly. This guide explains why the Wayback Machine returns a 429, how to clear it in a browser and in download tools, and how to plan a download so you do not trigger it again.

What a Wayback Machine 429 error actually means

HTTP status 429 is a standard response code that a server sends when a client exceeds a rate limit. The Internet Archive runs the Wayback Machine as a free public service on its own infrastructure, and it protects that service from heavy automated traffic. When a client crosses the line, the archive answers with 429 instead of the page you asked for.

Three points are worth knowing before you start troubleshooting:

• The limit is applied to you, not to the page. The same capture will usually load normally for someone else, or for you later.
• The exact thresholds are not published. The Internet Archive does not document a fixed number of requests per minute for replay or for its APIs, and it has changed its protections over time. Third-party guides quote figures, but treat any number you read as an estimate, not a rule.
• Ignoring 429 responses makes it worse. Tools that keep retrying immediately can turn a short pause into a longer block for your IP address.

Reports from 2026 describe the archive adding stronger protections against waves of high-volume automated traffic, and acknowledging that these can sometimes catch ordinary users too. That is one reason many people have started seeing the error recently even when they were only browsing.

Common causes of a Wayback Machine 429

Almost every case falls into one of these patterns:

1. A downloader running too many parallel requests. Website downloaders often fetch many files at once to save time. Against the Wayback Machine this is the fastest way to get rate-limited.
2. Retries without a pause. A script that retries a failed URL straight away sends a burst of identical requests, which looks like abuse.
3. Repeated CDX API queries. Listing every capture of a large domain, page by page, with no delay between pages, adds up quickly.
4. Save Page Now in quick succession. Submitting many pages for archiving one after another has its own, separate limits.
5. A shared IP address. On office networks, university networks, mobile carriers, VPNs and cloud servers, many people share one public IP. Someone else’s heavy traffic can put you over the limit.
6. Browser extensions or prefetching. Extensions that check archive availability for every page you visit, or that preload links, send requests you never see.

How to fix a 429 in your browser

If you only see the error while browsing captures by hand, the fix is usually quick:

1. Stop reloading. Every refresh is another request against the limit.
2. Wait. Give it a few minutes. If you were blocked after heavy use, it can take longer, sometimes an hour or more.
3. Close other archive tabs and pause scripts on the same network. Check whether a download tool, a colleague’s script or a scheduled job is running from the same connection.
4. Disable archive-related extensions for the moment, and turn off any VPN or proxy that might be shared with heavy users.
5. Load one capture, then move slowly. Opening a calendar view, then a capture, then its assets is several requests. Browse at a human pace.

If the error persists for a long time even though you have done nothing unusual, the Internet Archive’s help pages list how to contact them. A short, factual message with the time, your approximate location and what you were doing is the right approach if you think you were blocked by mistake.

How to fix a Wayback Machine 429 in download tools

Most people hit the error while downloading an archived website. Whatever tool you use, the cure is the same principle: one request at a time, with a pause, and a sensible response to errors.

Reduce concurrency to one

If your tool has a concurrency, threads or parallel downloads setting, set it to 1. For the open-source wayback-machine-downloader on GitHub, that means leaving the concurrency option at its lowest value rather than raising it to speed things up. For wget, avoid running several copies at once.

Add a delay between requests

Add a fixed wait plus some randomness. With wget, options such as --wait=2 --random-wait space requests out. In a script, sleep between requests and do not fire retries instantly.

Back off when you see a 429

Treat a 429 as an instruction to slow down, not as a failed file to retry immediately. A simple pattern is exponential backoff: wait 30 seconds after the first 429, then 60, then 120, and stop the job if it continues. If the response includes a Retry-After header, honour it.

Make the job resumable

Log every URL you have saved and every URL that failed. When the block clears, restart from the failures instead of downloading the whole site again. Re-downloading files you already have only spends more of your allowance.

Request less in the first place

• Use the CDX API with collapse=digest or collapse=urlkey so you do not fetch identical copies of the same file many times.
• Limit the date range to the snapshot period you actually need.
• Filter by status code so you skip redirects and error captures.
• Request original files with the id_ form of the URL, which returns the stored file without the archive toolbar and rewritten links, so you need fewer clean-up passes later.

Our guide to the Wayback Machine API and CDX queries shows how to build that list of URLs before you download anything.

A polite download plan that avoids rate limits

For a small or medium website, this order of work keeps requests low and makes 429 errors rare:

1. Check what exists. Run one CDX query for the domain to see how many captured URLs there are and which years are covered.
2. Pick the snapshot window. Choose the period where the pages you care about were captured completely. Our article on choosing the right Wayback snapshot explains how to judge this.
3. Build a de-duplicated URL list. One line per file, using id_ URLs.
4. Download slowly, one at a time. Leave the job running rather than trying to finish in minutes.
5. Log and retry failures later. A second pass after a break usually clears most of them.
6. Fix links and test offline. Only then is the copy worth packaging.

A site with a few hundred files can still take a while at a gentle pace. That is normal. Speed is the one thing the archive does not want you to optimise.

How to tell a 429 from other Wayback errors

Not every failure is a rate limit. Check the status code before you change your settings:

What you see	Likely meaning	What to do
429 Too Many Requests	Rate limit on your client or IP	Stop, wait, slow down, back off
404 or “Wayback Machine has not archived that URL”	No capture of that exact URL	Try other URL variants (www, https, trailing slash) or another date
Capture loads but shows an error page	The crawler saved an error the original site returned	Choose a different capture date
5xx errors or very slow responses	Archive-side load or maintenance	Wait and retry later; this is not your settings
Connection resets or timeouts	Network issues, sometimes linked to blocking	Pause, then test one request by hand

If your download is failing for reasons other than rate limits, our checklist for when a Wayback Machine downloader is not working covers empty pages, missing images, broken links and installation problems.

Save Page Now and 429 errors

Saving pages into the archive is a different service from reading captures, and it has its own limits. If you are submitting a list of URLs to Save Page Now, space the submissions out and do not run several at once. A burst of saves can block further saves from your IP for a period even while normal browsing still works. If you need to preserve many pages, spread the work over hours or days, and consider whether every URL really needs a fresh capture.

When rate limits mean it is time to stop doing it yourself

A 429 is a speed problem, and you can always solve it with patience. Where people get stuck is the combination: a large site, a tight deadline, captures spread across several years, missing images, and links that still point at web.archive.org. At that point the rate limit is only one part of the job.

If you would rather hand that over, our Wayback website download service collects the captures at a pace the archive accepts, removes the archive code, fixes the links and delivers a clean HTML ZIP. If you are not sure the archive even holds enough of your site, start with a free archive check first. We cannot speed up the Internet Archive or recover files it never captured, and we will tell you plainly what is and is not there.

Frequently asked questions

How long does a Wayback Machine 429 block last?

There is no published figure. Light cases clear within minutes. If a tool kept sending requests after the first 429, the block can last much longer. Stop all traffic, wait, and then resume at a slower rate.

Will a VPN fix the Wayback Machine 429 error?

Sometimes it does the opposite. VPN exit addresses are shared by many users, so you may inherit someone else’s limit. Switching IP addresses to get around a block also works against the archive’s protections. Slowing down is the reliable fix.

Does a 429 mean my archived site is gone?

No. The error is about your request rate. The captures are still there. Once the limit clears, the same URLs will load again.

Can I download a whole archived site without hitting 429?

Usually, yes, if you use a de-duplicated list, one request at a time, delays between requests and backoff on errors. It takes longer, but it finishes.

Is there an official rate limit I can configure my tool to?

The Internet Archive does not publish one for general use, and its protections change. Design your tool to respond to 429 responses rather than to a fixed number.