Launch HN: Cosmic JS (YC W19) - API-first drop-in replacement for WordPress
I'm Tony, one of the cofounders of Cosmic JS (YC W19) (https://cosmicjs.com). Cosmic JS is a drop-in replacement for WordPress that can power content for any website or app. We provide a web dashboard to create content and API tools and resources (REST and GraphQL) to integrate content into any new or existing project. Commonly referred to as a "Headless CMS", this eliminates the need to build and maintain your own CMS infrastructure. For a monthly fee, you use our CMS infrastructure and can focus on what really matters: building great products and user experiences.
My cofounder Carson and I met at a digital agency where we built and managed WordPress websites. We noticed that lots of development time was spent building and maintaining the CMS itself, sucking time away from core application development. Plus we encountered the same CMS problems over and over: automatic updates caused sites to crash, a client would decide to install a bunch of plugins that caused the site to crash, comment spam was a never-ending battle. We began looking for a better way to manage content.
This was 2014 and API services were becoming more popular (Stripe, Twilio, SendGrid etc were gaining traction in offloading non-core dev tasks), and it made sense that using an API could be a viable way to deal with content as well. So Cosmic JS was created to be the solution that we wanted to use: one click to add a new project, unlimited projects with a single login, a simple web dashboard to create content, and API tools and resources to integrate content into any new or existing website or app. No CMS infrastructure needed.
After much beta testing, we eventually released to the public in 2016. We're now powering production websites and apps for hundreds of teams around the world across various use-cases.
We know the market for a solution to this problem is big because WordPress, as of this posting, powers 30% of the web. That’s 75,000,000 websites (source: https://www.whoishostingthis.com/compare/wordpress/stats/). Plus the need for dynamic content extends beyond websites. Mobile, IOT and other emerging tech are increasingly requiring dynamic, easily integrated content.
This is a hard problem to solve because a CMS has to satisfy the needs of both developers and content creators. We're different than other headless CMS providers because lots of effort has been made to make the CMS admin dashboard and content integration process as easy as possible for both the developer and content creator. We’ve been told “it doesn’t get much simpler”. We’re also very committed to education and community. We're the only headless CMS that comes with a community of developers built-in providing hundreds of apps, extensions, and integrations to learn best practices and teach others. You can get up and running with a variety of use-cases in just a few clicks. And we have a free plan that rewards contributors with a free personal Bucket forever.
Check out some of the apps built with Cosmic JS: https://cosmicjs.com/apps.
We're excited to be participating in Y Combinator for the W19 batch to help more teams avoid the pain of CMS infrastructure management so they can focus on building great products.
We'd love to hear your feedback and learn more about your personal experiences building content-powered websites and apps!
92 comments
[ 3.1 ms ] story [ 118 ms ] threadAlso, terms such as "headless CMS" does not mean much to a regular user who is looking for a "site in WordPress". So is your intended audience primarily developers and freelancers/agencies ?
63.2% of developers voted it most dreaded in the 2018 Stack Overflow survey (narrowly beat by Windows Phone) https://insights.stackoverflow.com/survey/2018/#most-loved-d...
Our customers benefit by avoiding the pain of building and maintaining their own CMS infrastructure so they can focus on their core product. And we have a growing community that is adding more value through education and resources for every new user.
You're right, our target user is not a consumer who has many options for a consumer CMS. Our target customer is a team of developers and content creators.
The Wordpress ecosystem is mature and extensive. Wordpress developers are relatively inexpensive compared to frontend devs. Open source Wordpress is free as in beer.
> You're right, our target user is not a consumer who has many options for a consumer CMS. Our target customer is a team of developers and content creators.
Developers are expensive. Wordpress is not. A Twilio play ("Ask your developer" billboards) appealing to devs might not be as successful in this space.
I wish you well, but the headwinds are strong.
In the last year, WordPress grew from 29.9% of websites to 33.3%. Is that really losing favor? https://w3techs.com/technologies/history_overview/content_ma...
Unfortunately, this doesn't mean we want to switch. It still happens to be the best tool for the job 9 times out of 10, whether I hate it or not.
Or am i missing something?
Extensions are more like WP plugins: https://cosmicjs.com/extensions/
Thanks!
WordPress has problems. Big parts of WordPress sucks. But WordPress-the-thing is hard to argue with. It doesn't demonstrate a need to conflate sites with apps (maybe that impresses low-information folks but as a developer and a content creator I am left asking "so what?") and it doesn't need to hide its brain as a revenue model--they make a lot of money despite WP/WPMU both being open source. And somebody who has spent a decent chunk of time and money wrangling CMSes--I don't see a ton of value to a CMS I can't audit and can't just run myself. The `and` is important there. I'm happy to pay for support--and sometimes for operationalization, I've paid WPEngine a decent amount of money over the last few years--but I pay for support once I've validated the use case.
Perhaps I'm not the target audience. But I'm asking "why would I use this?" and coming up kind of empty. Beyond that, what does trouble me is the way in which you're intensely attempting to slag WordPress (which, tbh, rings hollow; "flies don't bite a man who's sure of himself" and all that). You can use WP as an API these days, and it's not bad. Not perfect, but not bad. But you can also do that with other headless CMSes, and I'm not really catching much differentiation between CosmicJS and those, either? The whole thing is coming off like low-information-targeting to me and that immediately makes me ask "OK, where's the grift?". I assume that isn't your intent, and I'm trying to give you the benefit of the doubt. But that's kind of the message I'm picking up here, and that might be something to work on.
And I spent years building WP websites, made a career out of it for a while, so it's not anything against WP and certainly don't intend to come across the way you described. So really Cosmic JS was built to satisfy my own needs that I didn't get from an installed CMS.
Wordpress is "free" vs CosmicJS is 49$/mo minimum without backup and only 3GB space.
For 49$/mo you could get a cloud VPS for Wordpress, including backups and even high-availability with no request limits (API limits in CosmicJS).
I know it's not straight comparable, but..
I mean, I see your point. I’m just not convinced it’s enough given the availability of a thousand other (free-ish) solutions.
Seems like one or two good fulltime developers can maintain a project like this.
You guys are a solution in desperate need of a problem.
I'm all for the headless CMS space, but I'll never recommend a solution to my clients that causes vendor lock-in and that I don't have the ability to operationalize myself.
Most folks in this space, wordpress included, follow a model of "open source, but pay us to scale and operationalize it for you".
They follow this model because it works.
As a CTO making decisions for my company, I'm not going to take the risk on a startup in this space without an open source fallback plan.
We have dealt with this a couple times with shopping cart providers, mls idx search systems, do-it-yourself website systems, captcha systems, cms systems (paid-for, now with no upgrade path for things like PHP5 to 7), payment processing systems, and more I probably can't think of at the moment.
If it's not open source, and self-hosted, then we walk away.
The only things we kind-of lock into now are hosting providers, email providers, and domain registrars. But we use pretty big and established operations for these just so we hopefully don't have to move things.
Hosting... Don't do some crazy VPS thing, you don't want to worry about being a sysadmin. Dreamhost has a good WP setup with most things you'd need ready to go. There are lots of other "managed WordPress" companies too.
Email, just FastMail or G-Suite (Google Apps). Probably just the latter because it's familiar to most.
Registration is pretty negligible these days. I like Namecheap and their corporate values and support.
Does anyone know how they made their demo thingee on the front page? It's super slick.
However, I realised that it is going to be bad for SEO as I cannot do serverside rendering this way.
I cosidered adding a WordPress blog, but eventually just ended up with a simple markdown rendering from the files - 1 file - 1 post.
If you manage to find a easy way to support serverside rendering, the product is going to be good.
Here's an article I wrote about opening up a blog with Cosmic back in 2017. It still holds... https://hackernoon.com/language-agnostic-content-management-...
> Cosmic JS is a drop-in replacement for WordPress
I don't know what "drop-in replacement" means in this context. To me, a drop-in replacement for WordPress would offer one or more of the following features:
1. Connect to my existing WordPress content database and use it as its content store
2. Use my existing WordPress theme to define its presentation
3. Provide APIs that are 100% identical to the existing WordPress APIs (the Plugin API, the REST API, etc.)
4. Allow loading and running existing WordPress plugins without modification
I don't see how Cosmic JS does any of these things. Some of them it addresses partially, an example being that there is a content importer available to pull content in from a WP site. But that still doesn't feel like a "drop-in" solution to me. A drop-in solution means I can pull out WordPress and replace it with the new thing without having to think much or rebuild anything, which does not seem to be the case here.
> We noticed that lots of development time was spent building and maintaining the CMS itself, sucking time away from core application development.
I don't understand this. WordPress famously installs in five minutes, and once that's done you only have to deal with it by installing updates, which is increasingly something it can do on its own without your intervention.
It may just be a matter of semantics - to me, "core application development" on WP is everything you do beyond installing WP itself: designing content architecture, selecting or building plugins and themes, etc. The "application" in the WP context is the complete bundle of database content, first-party code and third-party code you assemble to create a given site.
> automatic updates caused sites to crash
I have literally never seen this. Never. If anything, my complaint on automatic updates is that the WP core team has been so cautious to avoid breaking sites that it's made getting automatic updates into the hands of WP users a glacially slow process.
> a client would decide to install a bunch of plugins that caused the site to crash
Any sufficiently popular CMS that allows third-party extensions is going to have this problem. There are going to be third parties out there writing crappy code, and users out there who get dazzled by the marketing of that crappy code and install it. If your solution to this problem is to tell users they can't install third-party extensions, or to drastically limit what those extensions can do, you're going to be at a severe marketing disadvantage to systems that don't have those limitations. There are lots of dumb people out there who desperately want to do these dumb things, and don't like being told the reason they can't is because it's for their own good.
> comment spam was a never-ending battle
Turn comments off (which you'd do on any non-blog site you're building with WordPress), problem solved. If for some brain-damaged reason you want comments, install Akismet (https://akismet.com/), problem solved.
> We're the only headless CMS that comes with a community of developers built-in
WordPress itself is becoming a headless CMS. WordPress.com already provides a REST API (see https://developer.wordpress.com/docs/api/), so if you just want to host your frontend and let a remote service host the software and database, you can do that with WordPress. (You can also do it with self-hosted WP, but the value proposition is less obvious there since you still have to run and maint...
As for the semantics, you are correct, we are not proposing that you can use Cosmic to connect to your WP backend. We provide a light-weight solution to add dynamic content to any website or app. We offer a web dashboard to create content and API tools and resources to integrate content into any website or app. No infrastructure management needed. Our customers benefit from using Cosmic because they can avoid the pain of building and managing their own CMS infrastructure. But if you need that level of control and you like managing infrastructure, that's fine, then you may not be a customer.
We are seeing more and more services that offload these non-core infrastructure services: Stripe, Twilio, Algolia. They add value to development teams that do not want to manage infrastructure. That's our opinion on how content should be managed.
Hope this helps see things from our perspective. And I have taken your feedback in a cheerful spirit, as I hope you have taken my reply :)
Gutenberg promises to have WYSIWYG editable React components, which is a big deal, but they made insane decisions like storing the attributes in HTML, rendering HTML in the database, and requiring component developers to keep an array of deprecated changes when they want to modify anything on the component. In other words: you want to add a new CSS class to your component? Need an entirely new component for that in addition to maintaining the last one. Failure to deprecate will result in something like a fatal error in the editor that the user can't recover from. It's like PHP-era decisions with modern promises.
People are thrashing around in Github Issues, but the team seems unbothered. I would pay $$$$ for someone to unseat them.
This may be something we or our community may make available soon with enough interest.
Plus we're set apart in our commitment to our community and education. Our community contributes apps / extensions / open source code / learning center articles to help others with their specific use case. Our goal is for each new user to benefit from previous work in our community by offering pre-built apps, extensions and tutorials they can use to build a great product faster and easier.
At first I thought it meant I could maybe go to cosmicjs.com, enter a url and the username and password for my WordPress site, hit a button and have my site magically migrated to something that was 10x better.
But I guess what you actually mean is "WordPress clone" which doesn't seem as exciting.
>We're now powering production websites and apps for hundreds of teams around the world across various use-cases.
I guess I don't understand why would anyone join an accelerator with this level of traction. Seems that fundraising should be pretty straightforward.
I mean, you are providing a CMS and your own page is this slow?
Shouldn't speed be the main focus?
But c'mon... It's giving me 14 secs to interactive. How isn't that slow?
One thing that helped a ton was the availability of pre-built templates; there happened to be a node app template that fit my frontend use case perfectly, so I was able make the tweaks I wanted and launch.
First, a magazine startup with sloppy design when design should be their foremost asset (after content): https://news.ycombinator.com/item?id=19320608
Now a WP replacement which showcases and labels a sloppy editor as Delightful Experience ... https://cosmicjs.com/headless-cms#features--content-creators
Have you checked Gutenberg, the lates WP editor? It does blocks, like Contentful, Strapi and the other few dozens competing on the headless CMS scene.
Headless CMS is in equal part about content creation and content delivery ... a REST API and Graphql is not enough.
Sorry for the harsh words again but ... I do Wordpress since the beginning and I’m actively searching for a replacement. Yet dissatisfied again with a YC startup, the second disappointment in two days.
Tell me it’s coincidence.
I would hate for naive people to get trapped in your ecosystem and continue to get overcharged for no reason.
For the API requests per month pricing tiers, roughly how many API calls get made each time a site build is triggered for a site with say 100 blog posts?
Gatsby source plugin: https://www.npmjs.com/package/gatsby-source-cosmicjs
Gatsby Starter: https://github.com/cosmicjs/gatsby-starter
Gatsby App: https://cosmicjs.com/apps/gatsby-blog
Gridsome App: https://cosmicjs.com/apps/gridsome-blog
For API requests: To get all of your posts, you can use just one API request to the Objects endpoint.
Does this help?
1. A proper import / export tool for risk management
2. A slightly more refined message / call to action
3. Why the JS centric name/brand? Can it be used outside of NodeJS based tech stacks?
4. The price poice does seem high for small teams. You mentioned that this is normal for SaaS for some companies you listed. However, until you're established, it sounds a bit high. Can you perhaps go for a discount for the first X amount to time to entice potential customers?
Best! - K
1. We do offer a full data export. After you login, go to Bucket Settings > Import / Export to get a full JSON file export of your Bucket.
2. Thanks for the feedback. Anything specific?
3. Early on we wanted to be able to add dynamic content to any website server-side or client-side using NPM / React / Angular. And the domain name was available :)
4. Our target customer is teams, not individuals. And the value resonates even more with larger teams because they experience more of the pain. We've actually been told they would pay more :)
Especially in the Series A to Series B segment, where ability to change, and necessity to remove complexity, meet the sudden availability of funds to assist. That's where your example of "you can pay X and hour for your engineer or Y a month for us to handle it" works well. How will you pitch this moving further up-market? Especially where the opex exists that dev costs are less important. And vendor lock-in (or likelihood of existing in five years) becomes an issue?
And further up the enterprise scale, the greater the demand for a robust open source community. Is this on your roadmap? If not, why not? Especially given you use Wordpress as an example, and most of the industry itself using open source as a basis for content engineering tools. It's now the default expectation. Even moreso with the likes of near competitors like Strapi being open source. Thoughts?
Your app library is nice, but much like Gatsby or Hugo... it's not "wow". Any plans to ramp up the devrel? How will you stand out from all the other headless CMS products? Especially with the likes of Contentful so well funded. Are you worried that the lack of open source cuts off the lower end of the market and community growth, and the war chest of competitors at the higher level of the market constrains your customer base somewhat?
And.. wat will CosmisJS look like in two and five years? What I love about CMS teams is the creativity and intent of the awesome people who run them, so I'd love to hear more about the big vision too! :)
> How will you pitch this moving further up-market? Especially where the opex exists that dev costs are less important. And vendor lock-in (or likelihood of existing in five years) becomes an issue?
We believe that developer experience is table stakes now. It's becoming less the case that the CIO buys the tech and hands it to the development team. Developers are more and more making these technology decisions for their organizations. So it's our goal to empower the developers to help champion Cosmic for their team.
> And further up the enterprise scale, the greater the demand for a robust open source community. Is this on your roadmap?
Absolutely, that's why we're continuing to offer open source example apps, extensions, and integrations that can power complex business logic and to satisfy these types of customers.
> Any plans to ramp up the devrel?
Definitely, it's a big priority of ours to stay active in the developer community to listen and respond to use cases, technologies, extensions, and integrations that our customers will find valuable. This is our David / Goliath strategy. When you don't have a war chest to spend on marketing and sales people, you have to be resourceful :)
> And.. wat will CosmisJS look like in two and five years?
We've got an ambitious roadmap, so stay tuned! Simply put, our north star is to provide the most value for our customers by being the best solution in the era of API-first services.
You can use it to manage all variable things in your app, including content itself, of course.
It literally takes less than a minute to get started, because you can create an "anonymous" dropconfig without signing-up.
If you are looking for a self-hosted opensource MIT licensed option, checkout Daptin
https://github.com/daptin/daptin
I have been working on this since about a year and half, use it across a bunch of my own personal sites and desktop apps.
Would love to elaborate if there is interest :)