Just wanted to say that I love Micro.blog. It finally got me off my own Hugo plus voodoo setup because it’s so freaking ergonomic, and decently priced.
I had fun playing with SSGs for years. I’m having more fun just writing posts and letting them get broadcast to Mastodon on wherever else I’ve configured them at the same time. It wasn’t clear to me until I read Manton’s book, but its goal is to be a social media service that’s built completely on open web standards that everyone can participate with.
> Because if hosting videos were easy, YouTube wouldn’t be the only game in town.
Is self-hosting video still difficult, today in 2025?
My intuition is that there are less formats to worry about today, and serving video from static hosting that supports HTTP range headers may be enough for most devices to work.
What are the remaining hard problems? Maybe mechanisms to negotiate lower resolution for slower connections?
The hardest bit appears to be HLS - HTTP Live Streaming - the thing where a video gets divided up into lots of little .ts segment files and served via a m3u8 playlist.
I grew up in Australia, and more recently moved to India. Both are well away from the USA, where we who visit find the internet to be bafflingly fast just because of low latency (since most developers aren’t at all careful about avoiding request waterfalls, so even ignoring restricted bandwidth, 200ms of added latency makes the page load take several seconds longer).
Australia wasn’t great for playing high-bitrate videos hosted in the USA. With a high-quality 4G broadband link (rural western Victoria, clear line of sight to the tower 400m away that might host under 300 subscribers; at least 55/25Mbps via cheap phones when I measured it eight years ago), somehow you couldn’t actually download them at more than a few megabits per second at best. I think this is mostly latency effects, even though TCP is supposed to speed up.
India is terrible for playing high-bitrate videos hosted in the USA. A 100Mbps fibre connection can be completely undermined by what I imagine to be terrible peering arrangements on all broadband ISPs I’ve interacted with (Hireach, Airtel, Vybe), and files hosted in the US may trickle across at half a megabit per second or even less. Right now I can copy a file from my VPS in Australia at 1.6Mbps.
Hmm, I am Indian and I'd love to experiment with the setup, I have a 40 mbps fibre and I think my upload speed is very limited and I'd love if there are ways to replicate your setup? (maybe any websites which can do what you basically did and give me some stats which I can later share here?)
The situation between download and upload speeds isn’t likely to be much different, these days.
Speed tests (e.g. https://www.speedtest.net/) may let you choose a location to measure to. If I measure with https://www.speedtest.net/ to Planet Networks, Inc. in Newark, NJ, USA, it slowly climbs to 87Mbps/56Mbps. Trouble is, at greater distances these numbers are basically always unrealistic—you can’t actually reach them in normal usage.
Various hosts give you big files you can download to test them, which is often a more realistic measurement of what you’ll experience on the web at large. https://nj-us-ping.vultr.com/vultr.com.100MB.bin, for example, is from roughly the same geographical place, and downloading it in Firefox or curl only really reaches 1–1.3Mbps and would take 11–12 minutes. Whereas https://bom-in-ping.vultr.com/vultr.com.100MB.bin (Mumbai; I’m in Hyderabad) happily saturates my 100Mbps link and completes in nine seconds.
Depends on what exactly you mean with "self-hosting". As in you can run your own dedicated servers across the world, with one or more providers, and run all the software yourself, have your own CDN. Varnish works great for that. If you're limited to "self-hosting as in hosted in my home" then yeah it gets harder. Maybe P2P/P2P-like thing could help if you're dealing with public content, like PeerTube?
Sure, but CDNs for static content feel to me like effectively a solved problem in 2025.
There are plenty of good providers and some of them are practically free.
I mean sure, if you want to roll your own CDN by hosting boxes in colos across multiple continents and applying geographical load balancing via DNS you're taking on a whole lot of extra complexity, but I think outsourcing that to Cloudflare or Fastly or Fly.io or whomever is a reasonable strategy that still counts as "self-hosting", at least in comparison to using YouTube.
I want self-hosting to mean only hosting (as in, speaking HTTP) from servers that I control. I acknowledge this is not exclusively how the term is used any more, but I do think it’s still at least a preferable goal.
I want to self-host things. Currently I use a VPS. I’m planning on trying out hosting from home. Either way, if I get into doing much multi-megabit-per-second video stuff, hosting of that bit will definitely be going behind some CDN.
The other issue is that it's expensive. You can put a video on youtube for free and they carry the cost and cover it with advertising. If you self-host and your videos get a LOT of traffic it gets expensive quickly.
I feel people are often unreasonably scared by this sort of thing, so here are some numbers to give perspective and aid in decision-making.
Bunny CDN charge $5/TB for their volume network, which should be pretty good for video distribution, reducing after 500TB/month.
At a bitrate of 5Mbps (respectable for 1080p, significant overkill for more static types of content, as technical stuff will tend to be), 1TB is 444 hours. If, like OP, you publish 90-second videos, that’s 17,777 complete watches per terabyte. Depending on your situation, that might sound like not much or like a lot.
Put the other way round, at 5Mbps and $5/TB, each watch-hour costs $0.01125, a bit over one cent, and it takes 3,555 people watching your 90 second video to cost one cent.
For the sort of scale that most people are dealing with, it’s simply not an issue.
I don’t know if bots upset this balance. They may.
If you actually are spending more than a terabyte per month on it, then for technical audiences at least, I suspect that if you invited donations to specifically cover hosting costs (something along the lines of “I host these videos myself because ads and relying on YouTube are both bad for society; if you feel inclined, you can donate to help cover the cost, currently about $X/month”) you’d very quickly get a surplus. Or for longer-form content, charge something for 4K video (which costs 4.5¢ per watch-hour at 20Mbps and $5/TB) and let that subsidise the free 1080p (costing 1.125¢ per watch-hour) stream.
(On the $5/TB figure: my $5/month Vultr VPS includes 1TB per month, and charges $10/TB after that. Some VPS providers include a lot more; a Hetzner €3.49/month VPS in Europe includes 20TB then charges €1/TB. But remember, if you host video from one point only, that it is unlikely to work well for people halfway round the planet. See another of my comments in this thread for description.)
As for storage: each 90-second 5Mbps video is 56.25MB, and at a rate of $0.01/GB/month, each one will cost you $0.00675 per year to keep. Were you to post one 90-second video every single day and keep them all online, your monthly bill would grow by about $0.20 each year.
Another way to estimate upper/lower bounds for hosting videos is by considering the grey market adult video industry. It seems like there are hundreds, if not thousands of websites providing access to video, and from a unit economics perspective I remember reading those advertising CPMs are on the order of $0.01-$0.05 even for the biggest and least illegal websites like the Mindgeek properties. So I would assume the more shadowy websites operate at a budget of less than $0.01 per thousand views in revenue.
I’m assuming their CDNs are just specialised low cost hosting providers as opposed to p2p IoT botnets, but you never know.
I wish someone had the time and motivation to do an investigatory technical deep dive into the infrastructure they use.
LOT of traffic is a big if. A lot of videos only have 1k+ views and I will believe those are mostly drive-by viewing. Using a CDN can help you if you go viral.
I just setup private video for https://beautifulruby.com and can confirm the trickiest part is setting up and hosting HLS.
I extracted a RubyGem at https://github.com/beautifulruby/hls that I use to point at a folder full of videos, then my scripts converts them into HLS and uploads them to a private Tigris S3 bucket. I then have to rewrite playlists from the server with pre-signed S3 URLs.
It’s not that it’s difficult per se, but it does require a meticulous attention to detail to put all the pieces together.
I believe the bit where HLS is needed is when you want to change the format or the resolution on the fly. But browsers and othe players are perfectly capable of buffering and seeking single video files.
For most media, 1080p is fine enough. Add 720p and you have enough for 99% of the world.
It's a great idea and challenging YouTube's monopoly is noble, but I don't see how the economics work out. The current pricing [1] charges $20/month for videos up to 20 minutes of video, which is reasonable but still far too expensive for most people to use.
It's a great first step, but I struggle to see who this would be used by. So far Bluesky seems to be the only decentralized platform that's broken into the mainstream, and it'll only be more difficult for the video market.
Was curious about Cloudflare's pricing as a comparison... it looks like $1 per 1000 minutes viewed, which means a 10 minute video will cost #1 to have 100 views... That just seems prohibitively expensive to me, it's pretty much 100% of the lower end of video advertising just for delivery fees.
I know there are competing and cheaper services, but it still seems to be a big burden to get into. I've been trying to use Rumble a bit more, as well as appreciate the entry of Pepperbox, Floatplane and others. It's still a bit of a mess and none of them match the 10' experience of YouTube on Android TV, but it's getting better.
If you're not streaming live I believe you can serve video content out of R2 instead, which still somehow only charges for storage but offers completely free outbound bandwidth (egress).
Cloudflare Stream isn’t just for live-streaming, it’s for generic video hosting too. It does transcoding, adaptive bitrate, HLS, etc. It’s terrible and you shouldn’t use it, but it does a lot more than just serve static video files like R2.
Just to add, you should never serve video straight from R2 unless you pre-render all your videos for all major devices. Otherwise you get very, very subpar device buffering performance. Generally use R2 as a cache layer over your stream processor (CF Stream).
I may be misremembering, but I thought R2 terms disallowed video streaming content... I'm unable to find any references to this now, though I didn't do an exhaustive search.
> [...] Finally, we made it clear that customers can serve video and other large files using the CDN so long as that content is hosted by a Cloudflare service like Stream, Images, or R2.
Making a 20 minute long video is a sizable time investment. I imagine many people who are willing to invest that much effort into creating content would be happy to pay $20/month for hosting.
That's a fair point, especially considering everything else you get with the plan.
However, the pricing is still a far cry from non-decentralized solutions (for example, MUX's free plan offers 100K minutes per month [1]) and so the only other selling point is joining the fediverse—which is a good thing, but hard to get people to convert on (in Bluesky's case the turmoil that is now X was required).
Just wanted to say that I love Micro.blog. It finally got me off my own Hugo plus voodoo setup because it’s so freaking ergonomic, and decently priced.
I had fun playing with SSGs for years. I’m having more fun just writing posts and letting them get broadcast to Mastodon on wherever else I’ve configured them at the same time. It wasn’t clear to me until I read Manton’s book, but its goal is to be a social media service that’s built completely on open web standards that everyone can participate with.
> Because if hosting videos were easy, YouTube wouldn’t be the only game in town.
Is self-hosting video still difficult, today in 2025?
My intuition is that there are less formats to worry about today, and serving video from static hosting that supports HTTP range headers may be enough for most devices to work.
What are the remaining hard problems? Maybe mechanisms to negotiate lower resolution for slower connections?
UPDATE: Looks like this offers some answers to my questions: https://help.micro.blog/t/micro-blog-studio/4081
The hardest bit appears to be HLS - HTTP Live Streaming - the thing where a video gets divided up into lots of little .ts segment files and served via a m3u8 playlist.
Unfortunately, proximity can be important.
I grew up in Australia, and more recently moved to India. Both are well away from the USA, where we who visit find the internet to be bafflingly fast just because of low latency (since most developers aren’t at all careful about avoiding request waterfalls, so even ignoring restricted bandwidth, 200ms of added latency makes the page load take several seconds longer).
Australia wasn’t great for playing high-bitrate videos hosted in the USA. With a high-quality 4G broadband link (rural western Victoria, clear line of sight to the tower 400m away that might host under 300 subscribers; at least 55/25Mbps via cheap phones when I measured it eight years ago), somehow you couldn’t actually download them at more than a few megabits per second at best. I think this is mostly latency effects, even though TCP is supposed to speed up.
India is terrible for playing high-bitrate videos hosted in the USA. A 100Mbps fibre connection can be completely undermined by what I imagine to be terrible peering arrangements on all broadband ISPs I’ve interacted with (Hireach, Airtel, Vybe), and files hosted in the US may trickle across at half a megabit per second or even less. Right now I can copy a file from my VPS in Australia at 1.6Mbps.
And so CDNs are, very sadly, rather valuable.
Hmm, I am Indian and I'd love to experiment with the setup, I have a 40 mbps fibre and I think my upload speed is very limited and I'd love if there are ways to replicate your setup? (maybe any websites which can do what you basically did and give me some stats which I can later share here?)
The situation between download and upload speeds isn’t likely to be much different, these days.
Speed tests (e.g. https://www.speedtest.net/) may let you choose a location to measure to. If I measure with https://www.speedtest.net/ to Planet Networks, Inc. in Newark, NJ, USA, it slowly climbs to 87Mbps/56Mbps. Trouble is, at greater distances these numbers are basically always unrealistic—you can’t actually reach them in normal usage.
Various hosts give you big files you can download to test them, which is often a more realistic measurement of what you’ll experience on the web at large. https://nj-us-ping.vultr.com/vultr.com.100MB.bin, for example, is from roughly the same geographical place, and downloading it in Firefox or curl only really reaches 1–1.3Mbps and would take 11–12 minutes. Whereas https://bom-in-ping.vultr.com/vultr.com.100MB.bin (Mumbai; I’m in Hyderabad) happily saturates my 100Mbps link and completes in nine seconds.
That’s solved by CDN.
So it remains an unsolved problem for self-hosting.
Depends on what exactly you mean with "self-hosting". As in you can run your own dedicated servers across the world, with one or more providers, and run all the software yourself, have your own CDN. Varnish works great for that. If you're limited to "self-hosting as in hosted in my home" then yeah it gets harder. Maybe P2P/P2P-like thing could help if you're dealing with public content, like PeerTube?
Sure, but CDNs for static content feel to me like effectively a solved problem in 2025.
There are plenty of good providers and some of them are practically free.
I mean sure, if you want to roll your own CDN by hosting boxes in colos across multiple continents and applying geographical load balancing via DNS you're taking on a whole lot of extra complexity, but I think outsourcing that to Cloudflare or Fastly or Fly.io or whomever is a reasonable strategy that still counts as "self-hosting", at least in comparison to using YouTube.
I want self-hosting to mean only hosting (as in, speaking HTTP) from servers that I control. I acknowledge this is not exclusively how the term is used any more, but I do think it’s still at least a preferable goal.
I want to self-host things. Currently I use a VPS. I’m planning on trying out hosting from home. Either way, if I get into doing much multi-megabit-per-second video stuff, hosting of that bit will definitely be going behind some CDN.
The other issue is that it's expensive. You can put a video on youtube for free and they carry the cost and cover it with advertising. If you self-host and your videos get a LOT of traffic it gets expensive quickly.
I feel people are often unreasonably scared by this sort of thing, so here are some numbers to give perspective and aid in decision-making.
Bunny CDN charge $5/TB for their volume network, which should be pretty good for video distribution, reducing after 500TB/month.
At a bitrate of 5Mbps (respectable for 1080p, significant overkill for more static types of content, as technical stuff will tend to be), 1TB is 444 hours. If, like OP, you publish 90-second videos, that’s 17,777 complete watches per terabyte. Depending on your situation, that might sound like not much or like a lot.
Put the other way round, at 5Mbps and $5/TB, each watch-hour costs $0.01125, a bit over one cent, and it takes 3,555 people watching your 90 second video to cost one cent.
For the sort of scale that most people are dealing with, it’s simply not an issue.
I don’t know if bots upset this balance. They may.
If you actually are spending more than a terabyte per month on it, then for technical audiences at least, I suspect that if you invited donations to specifically cover hosting costs (something along the lines of “I host these videos myself because ads and relying on YouTube are both bad for society; if you feel inclined, you can donate to help cover the cost, currently about $X/month”) you’d very quickly get a surplus. Or for longer-form content, charge something for 4K video (which costs 4.5¢ per watch-hour at 20Mbps and $5/TB) and let that subsidise the free 1080p (costing 1.125¢ per watch-hour) stream.
(On the $5/TB figure: my $5/month Vultr VPS includes 1TB per month, and charges $10/TB after that. Some VPS providers include a lot more; a Hetzner €3.49/month VPS in Europe includes 20TB then charges €1/TB. But remember, if you host video from one point only, that it is unlikely to work well for people halfway round the planet. See another of my comments in this thread for description.)
As for storage: each 90-second 5Mbps video is 56.25MB, and at a rate of $0.01/GB/month, each one will cost you $0.00675 per year to keep. Were you to post one 90-second video every single day and keep them all online, your monthly bill would grow by about $0.20 each year.
Another way to estimate upper/lower bounds for hosting videos is by considering the grey market adult video industry. It seems like there are hundreds, if not thousands of websites providing access to video, and from a unit economics perspective I remember reading those advertising CPMs are on the order of $0.01-$0.05 even for the biggest and least illegal websites like the Mindgeek properties. So I would assume the more shadowy websites operate at a budget of less than $0.01 per thousand views in revenue.
I’m assuming their CDNs are just specialised low cost hosting providers as opposed to p2p IoT botnets, but you never know.
I wish someone had the time and motivation to do an investigatory technical deep dive into the infrastructure they use.
You're showing it's expensive to get a lot of views on a video, or as you say, it's not expensive if you get a small amount of views.
I want the fediverse and open web to be viable, and we need a solution to hosting high view count videos.
LOT of traffic is a big if. A lot of videos only have 1k+ views and I will believe those are mostly drive-by viewing. Using a CDN can help you if you go viral.
If we want the fediverse to take off we need a solution to the "LOT of traffic" issue.
A video CDN isn't cheap though, so outsourcing that to YouTube is the best option for most people.
Cloudflare egress is (somehow) free.
I just setup private video for https://beautifulruby.com and can confirm the trickiest part is setting up and hosting HLS.
I extracted a RubyGem at https://github.com/beautifulruby/hls that I use to point at a folder full of videos, then my scripts converts them into HLS and uploads them to a private Tigris S3 bucket. I then have to rewrite playlists from the server with pre-signed S3 URLs.
It’s not that it’s difficult per se, but it does require a meticulous attention to detail to put all the pieces together.
I believe the bit where HLS is needed is when you want to change the format or the resolution on the fly. But browsers and othe players are perfectly capable of buffering and seeking single video files.
For most media, 1080p is fine enough. Add 720p and you have enough for 99% of the world.
Why would you do this instead of just using html video tags?
The video tag needs something to link to, so you still have to solve the hosting problem somehow.
The weird m3u8 trick gets you better streaming and seeking performance, plus different video quality depending on the device and network connection.
It's a great idea and challenging YouTube's monopoly is noble, but I don't see how the economics work out. The current pricing [1] charges $20/month for videos up to 20 minutes of video, which is reasonable but still far too expensive for most people to use.
It's a great first step, but I struggle to see who this would be used by. So far Bluesky seems to be the only decentralized platform that's broken into the mainstream, and it'll only be more difficult for the video market.
- [1]: https://micro.blog/about/pricing
Was curious about Cloudflare's pricing as a comparison... it looks like $1 per 1000 minutes viewed, which means a 10 minute video will cost #1 to have 100 views... That just seems prohibitively expensive to me, it's pretty much 100% of the lower end of video advertising just for delivery fees.
I know there are competing and cheaper services, but it still seems to be a big burden to get into. I've been trying to use Rumble a bit more, as well as appreciate the entry of Pepperbox, Floatplane and others. It's still a bit of a mess and none of them match the 10' experience of YouTube on Android TV, but it's getting better.
$1/1,000 minutes is the pricing for Cloudflare's "Cloudflare Stream" product, which is specifically about live video streaming: https://developers.cloudflare.com/stream/pricing/
If you're not streaming live I believe you can serve video content out of R2 instead, which still somehow only charges for storage but offers completely free outbound bandwidth (egress).
Cloudflare Stream isn’t just for live-streaming, it’s for generic video hosting too. It does transcoding, adaptive bitrate, HLS, etc. It’s terrible and you shouldn’t use it, but it does a lot more than just serve static video files like R2.
Just to add, you should never serve video straight from R2 unless you pre-render all your videos for all major devices. Otherwise you get very, very subpar device buffering performance. Generally use R2 as a cache layer over your stream processor (CF Stream).
I may be misremembering, but I thought R2 terms disallowed video streaming content... I'm unable to find any references to this now, though I didn't do an exhaustive search.
Looks like they dropped that restriction in May 2023: https://blog.cloudflare.com/updated-tos/
> [...] Finally, we made it clear that customers can serve video and other large files using the CDN so long as that content is hosted by a Cloudflare service like Stream, Images, or R2.
Making a 20 minute long video is a sizable time investment. I imagine many people who are willing to invest that much effort into creating content would be happy to pay $20/month for hosting.
That's a fair point, especially considering everything else you get with the plan.
However, the pricing is still a far cry from non-decentralized solutions (for example, MUX's free plan offers 100K minutes per month [1]) and so the only other selling point is joining the fediverse—which is a good thing, but hard to get people to convert on (in Bluesky's case the turmoil that is now X was required).
- [1]: https://mux.com/pricing/video
It's also nowhere near long enough for most podcasts I've listened to.
Official announcement (that gets straight to the point): https://www.manton.org/2025/11/11/microblog-studio.html
Is that mean I can post "anything" included erotic related stuff like nude images and videos on it, and never worried about being deleted or banned?
Btw, indexing issue is another big problem on this platform.
This is a great little site that might motivate me to leave substack
[dead]
Now that Vimeo has been taken over by Bending Spoons it's good to have more hosting options.
[dead]