Open source alternatives to popular SaaS products | Hacker News Reader

Open source alternatives to popular SaaS products

github.com

299 pointscaaqil42 comments

Metabase is so fantastic - the guys who wrote it (originally called Caravel) met us in the AirBnB building when we were an early user - easily among my top 5 “most impressed” moments in the industry so far. I highly recommend it for absolutely anyone using SQL as their source of truth. Set up a read slave, host metabase, spend a week making pretty charts, make the C-suite love you forever. I’m pretty impressed to see this list from a VC firm!

While we’ve pivoted to more consumer friendly tools and less dev-tools (for now), I’m still working my butt off to make a future where self-hosted software (open source or not) can at least survive and ideally compete. I see three viable options: make it easier to install/manage in a users cloud (ala ReplicateD), make it easier for open source devs to host a paying user’s instance (ala rolling your own Stripe + AWS setup) and/or make it easier to physically ship preloaded plug-n-play hardware (my startup!)

If you’re an author of self-hosted software or anything on this list, feel free to reach out if you’re interested!

Some time ago there was a tweet making the rounds with a joke along the lines of:

"Congrats on your startup's Series A. Now you have enough runway to pay for 6 more months of

  - Jira
  - Github
  - Slack
  - Google Office
  - Google Cloud/AWS
  - LaunchDarkly
  - Contentful
  - Tableau
  - Zendesk
  - Salesforce
  - ...

I started to wonder if there is a market for a startup that simply takes all of the best-in-class open source alternatives and hosts the whole package that is commonly needed by companies. Would anyone here use something like that, and why don't smaller hosting providers invest in something like this? You see a lot of them offering basic Webmail/Calendar, but couldn't they go beyond that and use that as a possible revenue stream to compete with the big cloud providers?

you mean ... Startup-as-a-Service ?

Pretty much, yes!

For my own projects, I have:

  - Gitea
  - Drone/Harness CI
  - Taiga (issue tracker, project management)
  - Minio for Storage/Static site hosting (project site/blog/documentation)
  - Synapse (Matrix) for group chat, bots, video conferences.
  - Discourse (for support/community forum)
  - Redash
  - Syncthing for documents/assets, and basic Desktop apps (LibreOffice/Gimp/Inkscape) to edit them.
  - Wagtail for CMS (mostly Django projects), but also used Strapi before and it worked well.

I also was just this week experimenting with Baserow (https://baserow.io) to leverage a no-code platform to build feedback-forms, landing pages with contact forms, etc. Not 100% ready for it, but I'd guess pretty soon it will be.

Anyway, it's mostly me nowadays for these projects, but if lightning strikes and I need to grow a team, it wouldn't be hard for me to add Authelia with an LDAP server to provide SSO.

The big question for me would be (a) if I am too peculiar with this setup to the point that other people would not accept using these tools and (b) if the joke was just an exaggeration or if there is actually a market for funded startups that would be interested in saving some bucks per month by using F/OSS alternatives.

> Google Cloud/AWS

I mean, this is just operations. Unless startups start investing their entire series A into hardware and colocation contracts, they gotta have somewhere to compute their code.

> You see a lot of them offering basic Webmail/Calendar, but couldn't they go beyond that and use that as a possible revenue stream to compete with the big cloud providers?

That's because webmail and calendar are extremely easy to run with off-the-shelf software and a small team maintaining patches for it. As soon as you get into document collaboration, video calls, data retention/data loss prevention, SAML/LDAP, offering PBs of storage, etc. you end up needing a huge engineering team for each of those respective products, especially if you don't want to become stagnant on features and lose out on the value front.

> investing their entire series A into hardware and colocation contracts

I have the feeling for 80% of series-A startups, their entire client base could be served from 3 recycled notebooks locked in a janitor's room.

> Unless startups start investing their entire series A into hardware and colocation contracts,

Most startups can run on a $10/mo vps. Probably many fully grown companies can get away with not much more. The ‘everything cloud’ thing is popular and many investors see it as a sign the cto is an expert when the hosting bills for cloud hosting is high as possible, so I understand why it happens a lot.

I'm nervous about straying from Django ecosystem for its productivity. And people smarter than me using Rails have made peace with their sky high server bills. Has this changed?

That was never a thing. I had django websites running on 2GB of RAM at webfaction (shared hosting) with databases that got techcrunched and went on fine. I had sites with Django/Celery/Postgresql on a single server with 8GB RAM which worked fine for years.

Unless you have an unbelievably I/O or CPU heavy application, the tech stack is not going to be your bottleneck until you are in the tens of thousands of concurrent users.

Yep. One of our most used products is a Django site and it’s been on a 2gb server with MySQL and celery since 2013. No problems at all. Also 0 downtime, literally. Sometimes you get lucky!

Yes. https://elest.io/

It was on HN 4 days ago:

https://news.ycombinator.com/item?id=30440486

Thanks for bringing it up, somehow I missed that!

It is very close to what I have in mind, and I am glad to see they are also committing to share revenue with upstream developers.

I have heard of sandstorm.io

Communications from these platforms look so absolutely unprofessional. Having gone through a dialog with a couple familiar names now publicly traded it's strange watching the amateur hour that is this train wreck of dated, half-baked tooling. You can't tell if you're submitting personal, sensitive info to the business entity or if it's some phishing scheme.

i haven't got around to trying it yet but yunohost seems to solve part of this problem. it even helps you manage users. not every app is on there though

Take a look at https://www.restack.io/. BTW, I created this awesome list :)

I really like it. Are you in Berlin? I am working on something that might be aligned with your USP, I would be interested in chatting with you guys.

For more software alternatives, try AlternativeTo (crowd-sourced) or LibHunt (mix of crowd-source and ai).

[0] https://alternativeto.net/

[1] https://www.libhunt.com/

As the person building LibHunt, I hope it's helpful to you and thank you for sharing it.

Also, if you are looking for generic alternatives to popular SaaS solutions, I'd like to recommend SaaSHub https://www.saashub.com (another project of mine that is a direct competitor to AlternativeT). It's again a mix of crowd-sourcing and ML.

LibHunt was an instant bookmark for me when I came across it. The site is so well put together and cleanly designed, it's a pleasure to use.

I'm curious. I've been using LibHunt more and more but I find that the recommendations fall off in relevance after the first few recommendations. For example, if I go to https://www.libhunt.com/r/react-hook-form, the first two are relevant, the sponsored link isn't relevant, and then the projects are that after either full on app state solutions or UI component frameworks (admittedly, antd does have a useForm hook). How do your crowd-sourcing / ML models determine the relevance in a case like this?

I cannot seem to find a FOSS version of what I know as industrial / plant process monitoring. Think Input A and B go to Process C which is making Amount N and passing them to output D. Usually you would see this as a series of pipes, valves, machines, etc.

Maybe something like a SCADA visualization? IT would be excellent if it could be web based so I could share the view, and then feed it data from a back end db or similar.

What I'm actually interested in it for: Network monitoring for natural disaster information in connected systems. That way I can display how a series of rivers is being tied together, work in things like flow rate, topping rates for dams, alert levels, etc. Then, when things are getting hairy, issue alerts and change upstream/downstream colours.

Edit: Scadavis.io + node-red looks possible, but also looks like a complete pain.

If you find something less painful, poke me! I want this for systems of connected software.

Will do, and likewise!

How about an open source SCADA? I think Ignition by Inductive Automation is OS.

I might be missing something, but their pricing page suggests its quite an expensive bit of kit for something that's just a thought bubble currently.

Looks slick though, and supports MQTT

Another one for the No-code database category. My new favorite, heard about it on HN a few weeks ago - grist!

https://github.com/gristlabs/grist-core

Note that it's open-core. It's fine if it's not a problem for you.

When this project made it to the front page of HackerNews, they (grist-core owners) mentioned most things were open-source. Their process flow (which involves Phabricator and a mono-repo) hides a lot of the development activity but based on the repos commit history, they are putting effort into the repo as the following shows for the last 12 months:

      month  | authors | commits | files | churn |                                                          
    ---------+---------+---------+-------+-------+----------------------------------------------------
     2022-02 |      12 |      81 |   192 |  4951 | typescript:130,python:34,javascript:11,unknown:6
     2022-01 |       7 |      66 |   124 |  5380 | typescript:75,svg:14,javascript:9,python:9
     2021-12 |       6 |      45 |   115 |  3563 | typescript:85,javascript:9,python:7,unknown:7
     2021-11 |       6 |      69 |   133 |  5845 | typescript:95,python:18,javascript:5,svg:4
     2021-10 |       6 |      58 |   142 |  7718 | typescript:116,python:10,javascript:4,unknown:4
     2021-09 |       6 |      44 |   102 |  4223 | typescript:70,python:25,javascript:4,unknown:2,css:1
     2021-08 |       6 |      80 |   182 |  5425 | typescript:125,python:28,unknown:14,javascript:9
     2021-07 |       6 |      83 |   198 | 12876 | typescript:123,python:40,unknown:17,javascript:8
     2021-06 |       6 |      60 |   166 |  5488 | python:79,typescript:62,javascript:14,css:8,text:1
     2021-05 |       6 |      52 |   122 |  4846 | typescript:89,python:11,javascript:10,css:5,json:2
     2021-04 |       4 |      35 |    94 |  8006 | typescript:72,javascript:8,json:4,unknown:3,yaml:2
     2021-03 |       3 |      35 |    98 |  4361 | typescript:79,javascript:10,python:5,css:2,json:1

Note, I've truncated the file types changed to reduce the line width. Also, the number beside the file type is the number of files changed. For example, typescript:130 means 130 typescript files were changed.

how did you compute this table? with some kind of library/tool?

My product (https://gitsense.com) moves most of Git's history into a relational database and then I do indexing on top of that to produce things like code churn (lines added, changed and deleted), which is also stored in the database. So as you can probably imagine, generating the table is just a bunch of SQL statements.

im really liking grist. ive never worked with databases before so some things are still not clicking with me yet but being able to use python is such a killer feature compared to airtable

Hosting this is still a major hurdle for ordinary folks.

That's why I'm currently building a light and low-cost container hosting service to run selected FOSS apps. Setup and hosting is all taken care of, while still giving you full data access over SFTP if you want that.

Part of the revenue is also shared with the original authors to keep the project sustainable. Or donated to a cause of the author's choosing maybe.

Check out the free beta here: https://pikapods.com

I'm surprised office software are not mentioned. To Google Docs Editors / Office 365, an alternative could be ONLYOFFICE. Their word processor, spreadsheet etc work both online and as desktop software, high compatibility with MS formats, and they have a free personal tier and also business plans. I'm not affiliated, just a happy user. And they're EU based!

https://www.onlyoffice.com

EDIT: I see that the repo lists companies established "in the last 10 years" and OO's company was founded in 2009, so they're not a candidate for inclusion.

I'm sure that there are many amazing products on this list but I just want to give a shout out to Typesense (Algolia alternative) from personal experience.

The product is great and the founders are incredibly attentive to their community.

Posthog[1] is an excellent alternative to Google Analytics that should be in that list.

[1] https://github.com/PostHog/posthog

Posthog is in the list but in a different category.

Their 'can only use one project in the open source version' limit is extremely annoying though. They think people have only a single environment to deal with?

I don't like the single project limitation too but why does everyone have to provide a free version that you satisfy?

It's sad that Phacility shut down. The VCS browsing side of Phabricator, Diffusion, is a bit clunky compared to, say, GitLab (though GitLab locks several things behind a paywall, and clunky is better than nothing), but the issue tracker, Maniphest, is a competent offering.

This list makes me feel terrible. It's like I always use the other ones.

I also have some weird relationship with beauty, if the product is more beautiful I get attracted to it, forgetting some of the technical differences.

Reply on news.ycombinator.com