GitHub Community
-
Select Topic AreaQuestion BodyUsing Chrome browser, I visited a Release page: I chose a file "free-v10.3.0 Katalon.Studio.dmg" and tried to download it. The URL of the dmg file was The download continued approximatey 5minutes and terminated with failure message. I tried to download the file using wget command and curl command in the command line. I failed in both cases. In the wget session I saw a message: "error 618 jwt:expired". The following is the console output from wget: The following is the output from curl I tried downloading other files. I failed downloading files larger than approximately 300MB. I noticed that the session terminated after 5 minutes after the start. I have a connection to the Internet of 8Mbps --- ordinary speed, not too fast, not too slow, I think. It seemed small sized files are OK. I remember I could successfully download the similar-sized files last week. I guess that the JSON Web Token for GitHub Releases page has been changed recently: expiration was changed from a larger value to 5 minutes, which caused the |
Beta Was this translation helpful? Give feedback.
Replies: 29 comments 48 replies
-
Beta Was this translation helpful? Give feedback.
-
|
Thank you for your prompt reply.
When was this change made? I suppose, the expiration of JWT was shorted in the last few weeks, am I right? |
Beta Was this translation helpful? Give feedback.
-
|
that's not ideal - anyone who lives in an area that has poor or unreliable internet is then unable to download without this workaround. |
Beta Was this translation helpful? Give feedback.
-
Me for example.. Been trying to download a 1.5GB file and it always failed at around 300MB.. Now I know why... Security ffs |
Beta Was this translation helpful? Give feedback.
-
|
I tried GitHub CLI as suggested by Nabil-nl I wasn't successful. |
Beta Was this translation helpful? Give feedback.
-
|
I also tried downloading a resource using curl with my GitHub Personal Access Token: I wasn't successful. |
Beta Was this translation helpful? Give feedback.
-
|
Nabil-nil wrote:
To be honest, I don't understand what he suggested. |
Beta Was this translation helpful? Give feedback.
-
|
I'm also facing this error, I hope there will be a fix |
Beta Was this translation helpful? Give feedback.
-
|
The following issue discusses about the same thing: |
Beta Was this translation helpful? Give feedback.
-
|
Gift fake |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
-
|
What is written above is all correct and you can leave it for information, but it is better to remove the section with solutions because they do not work (and it seems you generated it with AI) |
Beta Was this translation helpful? Give feedback.
-
|
Such is the curse of bots. |
Beta Was this translation helpful? Give feedback.
-
|
Use wget's This is of course a work-around. I hope Github will stop this dropping of downloads after just a couple of minutes, it's preposterous. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you very much! This method really helped |
Beta Was this translation helpful? Give feedback.
-
|
@kazurayam Could you mark this answer as the solution to this problem? |
Beta Was this translation helpful? Give feedback.
-
I understand your point, but I would hesitate to mark this issue answered. By wget, by curl, by powershell ---- the choice does not matter. All these CLI tools are for the experienced programmers. But I remember I could download the 800MB .dmg file in a browser UI just by a single click on a HTML |
Beta Was this translation helpful? Give feedback.
-
|
Worked, size & checksum provided by project matched download. TYSM |
Beta Was this translation helpful? Give feedback.
-
This is worked for me, this is helping me so much, thank you for this |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
-
|
Please shut up, your answer is of no use If you want to help, share your solution, no one will read such long, pointless AI-generated ballads If we wanted to know how to get around this problem, we would all first ask AI because that's the easiest way, but as you can see, it didn't help and just gave useless advice, so we asked humans (which really helped) If you want to help, first check the methods suggested by AI yourself, if any really helped - then write here a short answer or a sample code that fixes this problem |
Beta Was this translation helpful? Give feedback.
-
|
+1 to @okineadev |
Beta Was this translation helpful? Give feedback.
-
|
It is 2025-08-13 21:04:26 in Japan; UTC+9 I retried downloading a large file from a Releases page using wget: It worked! No "jwt:expired" error occured. It seems that, just accidentally, I got connected to a GitHub access point with longer time-to-live token. |
Beta Was this translation helpful? Give feedback.
-
|
It is 2025-08-14 09:10:21 in Japan; UTC+9 I tried wget with -c option I failed: I got "jwt:expired" error. |
Beta Was this translation helpful? Give feedback.
-
You need to run exactly this command (which was provided in https://github.com/orgs/community/discussions/169381#discussioncomment-14082788): until wget -c https://github.com/katalon-studio/katalon-studio/releases/download/free-v10.3.0/Katalon.Studio.dmg; do :; doneWhen I ran this command to download my file, the download broke a few times (as expected), but it successfully resumed from where it left off, and for the first time I was able to download it. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you for your guidance. I made a bash script I tried the following. I could successfully download the file I wanted. |
Beta Was this translation helpful? Give feedback.
-
|
Still having this issue with the website on Firefox, |
Beta Was this translation helpful? Give feedback.
-
|
Github really needs to get on this. Guys; what if that was their primary source of downloads. They're depending on it working. I understand large files are a burden on the network no matter how you look at it but come on; with the amount of traffic you guys get that's minimal repo browsing surely you can afford to have a long-TTL token... |
Beta Was this translation helpful? Give feedback.
-
|
Large GitHub release assets now use short-lived signed URLs (JWTs) that expire in about 5 minutes, so your slow connection causes the download to fail once the token expires. |
Beta Was this translation helpful? Give feedback.
-
|
PowerShell has I haven't tried it; not going to do it myself. Please try. I hope you succeed and share the scriptlet here. |
Beta Was this translation helpful? Give feedback.
-
|
I checked the latest version of GitHub CLI. I checked if https://cli.github.com/manual/gh_release_download Unfortunately not yet. It's a shame of GitHub. |
Beta Was this translation helpful? Give feedback.
-
|
My bad, looks like Windows For Powershell, it requires version 7. To update, open up your current Powershell and enter: winget install --id Microsoft.Powershell --source wingetOnce you have Powershell 7, open it and use Invoke-WebRequest -uri https://github.com/ORG/REPO/releases/download/v1.0.0/release.zip -OutFile my_download.zip -MaximumRetryCount 20 -Resume |
Beta Was this translation helpful? Give feedback.
-
Good work! Thanks. |
Beta Was this translation helpful? Give feedback.
-
|
I didn't know of the "-C" option of curl. I checked the doc:
|
Beta Was this translation helpful? Give feedback.
-
|
When you click a release asset link (or use Historically, GitHub gave fairly long expirations (10–15 minutes+), but they’ve recently tightened it to ~5 minutes for some assets. Use That way, even if the token expires mid-download, Use a faster path or resumable download manager
(You still need to start from the fresh URL, because the token can’t be refreshed mid-download without a tool like Download via API with Authorization header If you use the Releases API with a Personal Access Token and (You can get But here is ready-to-run script Bash Script: Requirements
Example usage The script will:
Here is the parallel version with Script: Requirements Usage example How it works
|
Beta Was this translation helpful? Give feedback.
-
|
AI Generated |
Beta Was this translation helpful? Give feedback.
-
|
@kgoodwinurt wrote:
No, this is not true. When I tried the latest version of Apparently gh did not try to resume the session. |
Beta Was this translation helpful? Give feedback.
-
|
I didn't have success with I tried the second suggestion using aria2c which worked to download a 2gb file. I installed it with |
Beta Was this translation helpful? Give feedback.
-
|
Wow, this is what we're going to do to download a file... We're going backwards. |
Beta Was this translation helpful? Give feedback.
-
|
install.packages("remotes") 程序包‘remotes’打开成功，MD5和检查也通过 下载的二进制程序包在
Did you spell the repo owner (
Did you spell the repo owner (
Did you spell the repo owner (
|
Beta Was this translation helpful? Give feedback.
-
|
Esteemed Sir/Madam, |
Beta Was this translation helpful? Give feedback.
-
|
Turbo Download Manager (3rd edition) worked for me (FF) |
Beta Was this translation helpful? Give feedback.
-
|
Using third-party downloaders is not a good idea, I've never heard of it and I don't know on which platforms it will work and on which it won't And if, for example, my Bash script needs to download this file, I won't use third-party unknown tools instead of basic curl or wget So we need to know how to solve this problem using basic downloading methods |
Beta Was this translation helpful? Give feedback.
-
|
Turbo Download Manager : https://chromewebstore.google.com/detail/turbo-download-manager-3r/pabnknalmhfecdheflmcaehlepmhjlaa is a browser extention. You want to plug it into your Chrome or Firefox. It let me to download 800MB by a single operation on browser. You prefer a solution in the command line interface. That's OK. But I would say a solution in browser (without a Terminal window) is also needed. |
Beta Was this translation helpful? Give feedback.
-
|
Ran into the same error but in a CI/CD build script. Glad to hear it's not just me... Hope this gets fixed really soon, I was confused why downloading godot export templates would just error at 92% |
Beta Was this translation helpful? Give feedback.
-
|
I found this thread as I was having the same issue; but like many others in here (and other threads), it didn't quite help. I asked Copilot for help, and it provided the following PowerShell script for me, which worked great, so I thought I'd share it here. I asked it to write a bash version of the same script, but don't have anything to test it with, so let me know if it doesn't work for you. PowerShell version <#
.SYNOPSIS
Downloads a GitHub release asset in ranged chunks to avoid expiring signed URLs.
.PARAMETER ReleaseUrl
The GitHub release download URL (e.g. https://github.com/owner/repo/releases/download/tag/asset.zip).
.PARAMETER DownloadDir
Directory to save the file. Defaults to the folder where this script resides.
#>
param(
[Parameter(Mandatory)][string]$ReleaseUrl,
[string]$DownloadDir = $PSScriptRoot
)
# derive output path
$FileName = [System.IO.Path]::GetFileName($ReleaseUrl)
$Destination = Join-Path $DownloadDir $FileName
function Get-SignedUrl {
param([string]$Url)
$req = [Net.HttpWebRequest]::Create($Url)
$req.Method = 'HEAD'
$req.AllowAutoRedirect = $false
$resp = $req.GetResponse()
$loc = $resp.GetResponseHeader('Location')
$resp.Close()
if (-not $loc) { throw "Failed to get signed URL from $Url" }
return $loc
}
# 1) Figure out total size
$signedHead = Get-SignedUrl $ReleaseUrl
$headReq = [Net.HttpWebRequest]::Create($signedHead)
$headReq.Method = 'HEAD'
$headReq.AllowAutoRedirect = $false
$headResp = $headReq.GetResponse()
$totalSize = [int64]$headResp.GetResponseHeader('Content-Length')
$headResp.Close()
# 2) Open output stream
$fs = New-Object System.IO.FileStream(
$Destination,
[System.IO.FileMode]::Create,
[System.IO.FileAccess]::ReadWrite
)
# 3) Download chunks with progress
$chunkSizeMB = 5
$chunkSize = $chunkSizeMB * 1MB
$offset = 0
$startTime = Get-Date
while ($offset -lt $totalSize) {
$endByte = [math]::Min($offset + $chunkSize - 1, $totalSize - 1)
$signed = Get-SignedUrl $ReleaseUrl
$req = [Net.HttpWebRequest]::Create($signed)
$req.Method = 'GET'
$req.AllowAutoRedirect = $false
$req.AddRange($offset, $endByte)
$resp = $req.GetResponse()
$stream = $resp.GetResponseStream()
$buffer = New-Object byte[] 8192
$read = 0
$chunkDone = 0
$chunkTotal= $endByte - $offset + 1
# per‐chunk progress
while (($read = $stream.Read($buffer, 0, $buffer.Length)) -gt 0) {
$fs.Write($buffer, 0, $read)
$chunkDone += $read
$chunkPct = [math]::Round($chunkDone / $chunkTotal * 100, 1)
Write-Progress -Id 2 `
-Activity "Chunk bytes $offset–$endByte" `
-Status "$chunkPct% ($chunkDone/$chunkTotal bytes)" `
-PercentComplete $chunkPct
}
$stream.Close()
$resp.Close()
# overall progress
$offset += $chunkDone
$elapsed = (Get-Date) - $startTime
$speed = [math]::Round(($offset/1MB)/$elapsed.TotalSeconds, 2)
$remainSec = if ($speed -gt 0) { [math]::Round(($totalSize-$offset)/( $speed*1MB ), 0) } else { 0 }
$eta = (New-TimeSpan -Seconds $remainSec).ToString("hh\:mm\:ss")
$overallPct= [math]::Round($offset / $totalSize * 100, 1)
Write-Progress -Id 1 `
-Activity "Downloading $FileName" `
-Status "$overallPct% ($offset/$totalSize bytes) @ $speed MB/s – ETA $eta" `
-PercentComplete $overallPct
}
# 4) Cleanup
$fs.Close()
Write-Progress -Id 1 -Completed
Write-Progress -Id 2 -Completed
Write-Host "Download complete:`n $Destination"Example PowerShell usage and output: Bash version #!/usr/bin/env bash
set -euo pipefail
# parameters
release_url="$1"
download_dir="${2:-$(dirname "$0")}"
# derive output
file_name="${release_url##*/}"
destination="$download_dir/$file_name"
# chunk settings
chunk_mb=5
chunk_bytes=$((chunk_mb * 1024 * 1024))
# progress timers
start_ts=$(date +%s)
# helper: get signed URL from HEAD redirect
get_signed_url() {
curl -s -I -H "User-Agent: chunked-downloader" -X HEAD "$release_url" \
| awk '/^Location:/ {print $2; exit}' \
| tr -d $'\r'
}
# 1) get total size
signed=$(get_signed_url)
total_size=$(curl -s -I -H "User-Agent: chunked-downloader" -X HEAD "$signed" \
| awk '/^Content-Length:/ {print $2; exit}' \
| tr -d $'\r')
# ensure output exists
mkdir -p "$download_dir"
: > "$destination"
offset=0
while [ "$offset" -lt "$total_size" ]; do
end_byte=$(( offset + chunk_bytes - 1 ))
[ "$end_byte" -ge "$total_size" ] && end_byte=$(( total_size - 1 ))
# fresh signed URL
signed=$(get_signed_url)
# perform ranged fetch + write at offset
seek=$(( offset / 1048576 )) # in MiB
curl -s \
-H "User-Agent: chunked-downloader" \
-H "Range: bytes=$offset-$end_byte" \
"$signed" \
| dd of="$destination" bs=1M seek="$seek" conv=notrunc status=none
# update counters
bytes_read=$(( end_byte - offset + 1 ))
offset=$(( offset + bytes_read ))
# compute speed, ETA, percent
now_ts=$(date +%s)
elapsed=$(( now_ts - start_ts ))
speed=$(awk "BEGIN{printf \"%.2f\", ($offset/1024/1024)/($elapsed>0?elapsed:1)}")
remain_sec=$(awk "BEGIN{printf \"%d\", (($total_size-$offset)/1024/1024)/($speed>0?speed:1)}")
eta=$(date -u -d @"$remain_sec" +%H:%M:%S)
pct=$(awk "BEGIN{printf \"%.1f\", $offset*100/$total_size}")
printf "\rOverall: %s%% (%d/%d bytes) @ %s MB/s – ETA %s" \
"$pct" "$offset" "$total_size" "$speed" "$eta"
done
echo -e "\nDownload complete: $destination" |
Beta Was this translation helpful? Give feedback.
-
|
Bruh you can just use |
Beta Was this translation helpful? Give feedback.
-
|
Bruh, that, and no other solution I could find would work in my particular circumstance. I'm in a very rural area with spotty internet, and this was the only solution that worked for me, that is all. |
Beta Was this translation helpful? Give feedback.
-
|
I had the same issue. I used ADSL, AT&T 4G and my employer's network (a $90B+ corporation) and 5 different browsers. Failing on a browser could never be restarted, let alone provide a useful error. What worked was several iterations of |
Beta Was this translation helpful? Give feedback.
-
|
I think this is the easiest way to solve the problem – without third-party software, just using wget/curl from the console. It seems to me that Github wants to dictate a certain way of using the service to its users – these are inherent barriers to usability. Let's see where this leads. |
Beta Was this translation helpful? Give feedback.
-
|
💬 Your Product Feedback Has Been Submitted 🎉 Thank you for taking the time to share your insights with us! Your feedback is invaluable as we build a better GitHub experience for all our users. Here's what you can expect moving forward ⏩
Where to look to see what's shipping 👀
What you can do in the meantime 💻
As a member of the GitHub community, your participation is essential. While we can't promise that every suggestion will be implemented, we want to emphasize that your feedback is instrumental in guiding our decisions and priorities. Thank you once again for your contribution to making GitHub even better! We're grateful for your ongoing support and collaboration in shaping the future of our platform. ⭐ |
Beta Was this translation helpful? Give feedback.
This comment was marked as disruptive content.
This comment was marked as disruptive content.
-
|
I don't think it is a matter of "years and years". The download failures due to an apparent reduction of the token expiration period seems to be a quite recent change. Of course it is possible that even the previous more generous expiration could have cause some downloads of very large files on slow Internet connections to fail, but I am only seeing user reports of the problem from quite recently. Of course we are all hoping for something to be done to allow these downloads to be performed easily. I can't imagine what benefit GitHub would get from the expiration period change, but I'm also not knowledgeable in this area. It doesn't seem like a very effective way to reduce bandwidth usage, since people are going to unnecessarily attempt the failed download multiple times, actually using more bandwidth than if GitHub had just allowed it to go through. They could achieve that much more effectively through other means such as notifying repository owners if their usage is excessive, or imposing a file size limit on release assets. So I am doubtful that was their motivation for the change. |
Beta Was this translation helpful? Give feedback.
-
|
Various solutions for different CLIs: (sourced from comments here) Replace Powershell*: (majority of Windows users) Invoke-WebRequest -uri <url> -OutFile PLACEHOLDER.zip -MaximumRetryCount 20 -Resume*Requires Powershell 7. To update, do the command below before running the previous command: winget install --id Microsoft.Powershell --source wingetcurl: wget: aria2c |
Beta Was this translation helpful? Give feedback.
-
|
So we basically just need to set the -continue option for our downloaders |
Beta Was this translation helpful? Give feedback.
-
|
I tried downloading on Chrome and Firefox both on Windows and Fedora and even tried using mobile data in case it's my ISP. Always failed. Trying with ghTrying with curlAt 87% fails Trying with wgetSuddenly stops and leaves a weird output in the console at 87% |
Beta Was this translation helpful? Give feedback.
-
|
Nevermind. aria2 method worked. |
Beta Was this translation helpful? Give feedback.
-
|
This is still an issue!!!! This needs to get sorted out, not everyone has the damned luxury of fast internet, I am sitting in Africa with as fast a connection as I can get from my provider (50Mb) and am unable to download a frikken 1.2Gb file!!!!! Please, think of some other people other than yourselves sitting in other parts of the world !!! |
Beta Was this translation helpful? Give feedback.
-
|
upvoting; i'm in ontario, canada, rural. cant download files over about 750mb. |
Beta Was this translation helpful? Give feedback.
-
|
See my post with the best option we seem to have right now...maybe it will work with your 50Mb connection. Good luck! |
Beta Was this translation helpful? Give feedback.
-
Hi, thanks, I have tried ALL suggested options bar those for linux as I am on windows. The only one that got close was @SamMedo05 aria2. But. This still failed at 800Mb..... None of the softwares sped up any downloads for me. the through put on my card was only in the order of 1.5 to 4.8Mb at best. With a standard download Brave/ google/ firefox this was the same...... As good as Github is their downloads are for the dogs when it comes to larger files.... And yes. I am logged into my account when trying the downloads.... No difference.. |
Beta Was this translation helpful? Give feedback.
-
|
Yeah I tried almost all solutions above, the only one working for me was using (flags If the release is also available on a different host like SourceForge, downloading from there worked for me as well. |
Beta Was this translation helpful? Give feedback.
-
What ACTUALLY worked: aria2c -x 16 -s 16 "url-to-file" It sped up so much that it can download ~7GB before the 5-minute json thing expires and kills the download. WOOHOO |
Beta Was this translation helpful? Give feedback.
-
|
That may have been a temporary glitch in the matrix. The usual `` until wget -c ; do :; done '' worked fine this morning. We have the fastest ADSL available in Vermont but, with wifie's streaming, all I had to work with was 0.73 Mb/s. Started the dl, took breakfast, worked in the yard a bit, came back in and just like magic I had the file. Maybe someday they will pull fiber near the house---maybe 10 more years? |
Beta Was this translation helpful? Give feedback.
-
|
Like in my reply to you above this was a futile exercise for me... |
Beta Was this translation helpful? Give feedback.
-
|
On android worked with FDM downloader and main link from releases |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
-
|
GitHub support replied to me today:
Can someone test? |
Beta Was this translation helpful? Give feedback.
-
|
I tried downloading the following files (each 741 MB) from the web page in Chrome browser. I just clicked the link once per each. It seems that I no longer need wget -c. |
Beta Was this translation helpful? Give feedback.
-
|
Try a file bigger than a gig I linked in the file that does not download because of the certificates time constraintes, you tried with a file that can be downloaded in less than 5 minutes, try something that requires a longer time and report back... If you have realy fast internet access you will not have issues, if your internet is slow you are cooked. Stop saying you did it so everything is fine, the "if my download works then there is no problem" schtik is getting old quickly... Clearly, small files fall within the download certificates time allotment...... Try with a file like mine a file bigger than a gig. |
Beta Was this translation helpful? Give feedback.
-
The file size is not important, what matters is that the file should take a long time to load on your network. |
Beta Was this translation helpful? Give feedback.
-
I have tried the same file and it has downloaded now without giving any errors... My "internal network" or LAN runs at 2.5Gb but, my ISP only supplies 50MB per second... 
|
Beta Was this translation helpful? Give feedback.
-
|
I tested and found that I am now able to successfully perform a 12 minute release asset download without requiring any resumptions, even though the same download previously failed for me after 5 minutes. Thanks @okineadev for taking the time to post this valuable update! |
Beta Was this translation helpful? Give feedback.
and others
Turbo Download Manager (3rd edition) worked for me (FF)