Nest self hosting

is Nest available for self hosting? To be more precise: can one host private instance of Nest?

1 Like

Not yet. There are currently two very different versions of the Nest, of which the most recent one was supposed to become open source and self-hostable, but it never worked out. I’m working on merging their features. My main goal is that I want private repos for myself and a few friends.

As far as self-hosting is concerned, the entire Pijul project is currently unfunded, so I wouldn’t mind people self-hosting, as long as the public Nest pays its server bills and our engineering time, which is not currently the case.

But as soon as this works out, sure! I’m looking forward to it. If you know any way to help us get funded faster, please let me know.


Hi thank you for the answer.

Is ti great that you have it somewhere on the road map. Timelines and funding is hard so there is no expectation form me to have that available in near future. And exactly like you, I had idea to try out pijul with few friends and although ssh is option and it works it is convenient to have some minimal UI for repos like Nest.

In any case if I have some input on funding I’ll for sure get back to you as IMHO you are doing cool thing with pijul overall.

Quite often I see these types of posts about,, self hosting. This applies to zulip as well, it would be good ide to pin down post about this, because many people are asking this, many newbies are confused.

But as soon as this works out, sure! I’m looking forward to it. If you know any way to help us get funded faster, please let me know.

pmeunier, you should absolutely get a patronite, buymeacoffee, gofundme or other account. Pijul and Nest are large, ambitious and long-term projects. They need additional financing, especially since you are their only major developer here. You might be surprised what kind of financial support you get.


My next blog post will be about upcoming changes in the Nest, actually. I’ve decided to retire the serverless version, it’s slower than the kind of stuff I like to write.

And yes, why not a funding thing like that? I might actually try multiple routes at the same time.


When considering crowdsourcing platforms, I can also recommend taking a look at, which supports EU payment options by default.

I would also add a cryptocurrency wallet to the mix. Many big projects accept donation in crypto, like Internet Archive or even Free Software Foundation, so this is nothing uncommon. Just need to get it right. Many banks don’t like it when you exchange funds with entities dealing with crypto. Might block transactions or suspend your account for investigation, even terminate the contract for your banking services. So its better first familiarize with rules and regulations of your bank and your country.

Plus we are months from Halving, which could translate into sizable increases, which are already slowly beginning. So a 300% bonus would be nothing out of the ordinary. But this is just a theory. But this is nice thought indeed

Hi @levi and @Jummit, I just looked at OpenCollective and at getting paid in crypto. Since I already have a “fiscal host” according to the OpenCollective definition, I’m wondering what would prevent people from just subscribing to private accounts on the Nest rather than going through OpenCollective/BuyMeACoffee/… Maybe I should add an option to make a one-time donation? If any of yous can tell me what you expect or think others are expecting, that would be quite helpful.

About cryptos, it turns out the accounting required to accept payments in cryptos is quite complex, so it would have to be a donation to me as a person. Why not? Not sure how to go about transparency, but if you want to share any insight on how you’d like to see it happen, by all means do!

Hmmm…ok, so here is how I see this.

For me these are two different things. One thing is keeping business up and running, to pay for the hosting, domain and other related services and your bills. The second is support development of the project.

You can also try only one path first - the donation path. Because at this point Nest doesn’t offer much to charge money for. This could be hard. And it can also be daunting. Imho there is higher demand for additional features in Pijul than for the Nest. Maybe I’m wrong ? Private accounts will be rather scarce at first. Because there its more likely that new project and startups will pick up pijul as their new VCS, rather than well established projects with switch to Pijul. New project doesn’t have special requirements. But the well established have all kinds of needs to integrate their workflow with new tools. So it could take some time before people will need them, thus it will take some time before money could be charged and needs will be known.

I would do this at first. Start donations, and if you want to charge for the nest, then for “special” needs. Like additional CI, bigger quota, 24h backup. Since the nest doesn’t offer much yet, so it’s up to people to say what they need and what they would be able to pay for. When Nest will be mature enough, then some payment plan can be established. But then you will then have a better understanding of the market and users needs and expectations.

I would also collect targeted donations. For a specific thing. For example if you gather 5000$ you will start to work on supreme Git importer/exporter. Or 2000$ for binary patches, or per character patches (I have already seen that kind of request). another Idea is TUI, like LazyGit. Just examples. The community can vote on what it needs. You can post a new thread or a poll. Discourse can do that ? Ask the for features users find most lacking.

Choose one or two features and users will donate. The feature that reaches your designated target faster wins.
People love to support but they love even more if they can participate in direction the project it heading. Plus vote with money is much “heavier” that board post. You will know that people really need that feature. It’s also a good idea because it could attract companies with their specific needs. This also boost the project credibility.

Over time, you may need to separate yourself as a legal entity from your business by establishing a limited company or company that runs the Nest. Then you run the business and the donations will be only on you. But that is more distant future.

As for crypto, that train has already left. Check out what’s happening with BTC. The 300% predicted by me is slowly becoming a reality. It is now a very violent period for cryptocurrencies by the upcoming halving and new market cycle. It’s a tough subject right now.

Could the server less version be open sourced, perhaps we can help improve performance?

I like the idea of a server less implementation that can be deployed to Cloudflare

If you want to work on the performance of Sanakirja, you’re obviously more than welcome! Sanakirja already is the fastest open source KV store, but improvements are certainly possible and welcome, we’d definitely learn lots if any strong performance gain were to be found there. Maybe look at Patrascu-style datastructures and self-optimising trees?

If you want to stick to B Trees, the main issue is to run an allocator on top of Cloudflare KV. You don’t really need any specific code for that, other than the sanakirja-core crate (which compiles to WASM).

I have strong opinions about Cloudflare Workers’ performance and I believe I know how to do better (I even have prototypes). You could definitely start working on a new FaaS technology if you want, the main requirements is to allow its users to allocate replicated memory by blocks of 4kb, and be transactional (and ideally efficently forkable).

Know I’ve said this before, but there’s also a way to help effectively get funding without needing money, so to speak - more comments/documentation! I think a lot of the community would like to hack away and contribute, so maybe a couple of weekends just documenting more what you’ve already done - be this theory of operation or simply adding more inline comments about what each segment of code is doing and what it’s for might mean others could carry some more of the weight, over it being a lot on you Pièrre, and some other core developers. Having been a systems programmer for 15 years now, I still find in most projects where I have full access to the source code, I need to reverse-engineer what is happening. That slows down my ability to contribute massively, because I’m making second guesses and assumptions - so personally in my own code, I try really hard to spend a minute here and there describing what it is I was thinking when I wrote something, why it’s necessary, etc., so that when I look back later completely having forgotten it, the whole process is much faster and I don’t need to go through the process of figuring it all out a second time. 30-50% comments or newlines grouping things into paragraph is a pretty good thing to aim for IMO. I think it’d be worth a couple of weekends for pijul, and then the project may grow faster on its own.