Free Tiers Are Disappearing — What That Means for Developers
Free tiers are a well-worn marketing play: give users a taste of the product, let them settle in, and hope enough of them eventually convert to paying customers. But the model is increasingly under pressure. Service after service has sunset its free offering, and developers who built projects on those tiers are left scrambling. The disconnect is striking — everyone understands that products cost money to build and maintain, and most of us are happy to pay for tools we rely on. What grates is the bait-and-switch feeling of investing time and resources into something advertised as free, only to have it moved behind a paywall.
The annoyance is manageable when switching costs are low — if your favorite browser goes paid, you just find another. But for hosted products and services, migration isn't so trivial. Hundreds of thousands of developers may have deployed projects to a free hosting plan that suddenly has a price tag. The pattern is becoming distressingly common: companies retire free tiers once the product matures enough to stand on its own or when free customers fail to convert into paid ones at the expected rate.
Pricing Terminology Worth Knowing
Before digging into the fallout, it helps to establish a common vocabulary. "Free tier" actually covers several distinct models:
- Free trial opt-in — Users try the product for a limited period without supplying payment details. Access ends when the trial ends.
- Free trial opt-out — Payment information is collected at registration, and the free trial automatically converts to paid when it expires.
- Freemium model — Core features are free, while additional functionality requires a paid upgrade.
- Reverse trial model — Users begin with premium access and drop to the freemium tier when the trial period concludes.
Case Study: PlanetScale's Hobby Tier Removal
PlanetScale, founded in 2018, launched its database-as-a-service in 2021 and raised $105 million in venture funding. In March, CEO Sam Lambert announced the end of the company's hobby tier, framing the decision as a move toward "a reliable and sustainable platform for our customers" rather than "giving away endless amounts of free resources to keep growing." Free-tier users were given until April 8 to migrate or pay a starting price of $39 per month.
The messaging was contradictory. The memo warned that "[e]very unprofitable company has a date in the future where it could disappear", yet simultaneously touted that PlanetScale is "the main database for companies totaling more than $50B in market cap" and "one of the fastest growing tech companies in the US." Indeed, in November 2023, Deloitte ranked PlanetScale as the 188th fastest-growing company in North America — recognition that requires base-year operating revenues of at least $50,000 and current-year revenues of at least $5 million. Growth doesn't guarantee profitability, but the data suggests the company wasn't on the verge of collapse.
The decision reads more like a revenue play than a survival measure. That stings for the independent developers and bootstrappers who rely on free tiers to keep costs down, especially since the announcement coincided with layoffs. Even by industry standards, the pricing is steep — Heroku's entry point is $7 per month and Railway's is $5.
What a Free Tier Actually Buys the Provider
A new service with no users can't demonstrate its value behind a paywall. A freemium model is a common way to generate awareness and attract early adopters, with the hope that paying customers will subsidize everyone else. This is the old 80/20 assumption — roughly 20% of paying customers should cover the 80% who don't pay. The conversion rate, the percentage of users upgrading from free to paid, varies by model, but a good freemium number is around 3–5% even without sales assistance. Reality often lands lower, and startups with few customers may operate at a loss while trying to find enough paying users to even the books.
"
Once the audience is large enough, a company can afford to alienate free users to chase the minority who will pay. Evernote pulled this maneuver in 2023, slashing free limits from 100,000 notes and 250 notebooks down to just 50 notes and one notebook — a move that forced long-time free users to either upgrade or leave.
Free tiers serve different roles in SaaS versus PaaS, and the consequences of dropping them differ too. Cloud computing is expensive to provide, so free tiers impose real costs on PaaS providers in a way they don't for typical SaaS apps. Zoom offers a useful SaaS comparison: its basic tier provides 40-minute meetings, which suffices for many users. Remove that tier, and free users will simply migrate to Google Meet or another alternative — they've invested nothing that locks them in.
PaaS is another story. If your architecture lives in a provider's free tier, switching introduces real friction. Migrating data and servers costs money, partly due to data egress fees — charges cloud providers levy when customers move data off their platform. In effect, you pay to stop paying. The European Union's Data Act aims to curb this practice by requiring European cloud providers to remove barriers to switching, including artificial egress fees, but the problem persists elsewhere.
Trust and Transparency in the Freemium Compact
Who bears responsibility when a free tier disappears? There are two lenses. On the principle side, a company that markets itself as a reliable and sustainable platform shouldn't pull the rug out from under users who trusted it. On the consequential side, prudent developers should plan for the worst. You don't expect someone to run a red light, but you still check both directions before crossing.
The core issue is transparency. Free tiers are seldom labeled as temporary, even when they are. They're presented as a legitimate tier — fewer features than paid plans, but just as dependable. Users have no reason to expect their projects will need to migrate on short notice. As removals become more common, that assumption looks increasingly naive, but the onus shouldn't be on developers to anticipate corporate strategy when providers won't state their intentions upfront.
A Better Model Than The Free Tier
The practice of dangling a free tier in front of users only to remove it once the company has captured enough market share is fundamentally flawed, especially when there is no clear sunset date communicated at sign-up.
Pretending that the purpose of a free tier is the same as a free trial is unjust since it surely isn’t advertised that way.
If the goal is simply to give people a taste of what a product or service can do, several more honest approaches exist that don’t carry the same baggage as an indefinite free tier.
- Opt-in free trials. Strapi, the open-source CMS, is a solid case in point. When Strapi launched its cloud hosting service in 2023, it chose to offer a 14-day free trial without requiring a credit card up front. While Strapi Cloud isn’t cheap, the trial model is fair: it gives you the time to actually evaluate the product while avoiding the trap of forgotten subscriptions that some companies count on.
- Free credits. Railway is the best example I have seen of a company using credits to let users evaluate its platform. I’ve used it to host Node.js and Postgres projects, and the structure is straightforward: the base plan is a modest $5 per month, and new users receive $5 in credits to run their first project and kick the tires. No credit card required, and no rug to be pulled when the initial allocation is exhausted — you simply buy more credits if you want to keep going.
Railway is a particularly instructive case because it actually did discontinue its free tier on June 2, 2023, but the transition was handled with a level of care that PlanetScale failed to deliver. Railway gave its free-tier users a trial account with free credits, and its replacement pricing starts at only $5 per month. PlanetScale, by contrast, sunset its free tier and started charging $39 per month for its cheapest plan. The difference between a $5 and a $39 replacement cost is a matter of intent as much as arithmetic.
Free Tiers Worth Keeping
This conversation isn't meant to be a simple directory of free services, but rather a starting point for thinking about the actual value of the free-tier model. To that end, here are a few tiers that I rely on, even for small production-ready work.
Supabase
Supabase’s free tier is generous enough to build almost any imaginable back end. It combines a PostgreSQL database, authentication, real-time subscriptions, and file storage in one central dashboard, making it a solid foundation for a wide range of web apps.
Railway
Despite not having a permanent free tier, Railway remains my go-to for hosting Strapi CMS. Its strengths are polish and workflow: an attractive UI, seamless deployments, automatic scaling, CI/CD pipelines, and hundreds of templates for popular frameworks and databases. The $5 starting credit gives you the full ride, making it easy to evaluate without risk.
GitHub Pages
I use GitHub Pages in the classic way — for static pages, demos, and live code examples tied to blog posts. It’s a convenient playground for quick artifacts, but not something I depend on for anything consequential should it suffer an outage.
Netlify
Netlify does far more than just static hosting. It supports modern frameworks and bundles continuous deployment, documentation, templates, an edge network, and analytics all into a free tier that’s hard to beat.
The Cost Of Confusion
To be clear, the argument here is not that free tiers should be abolished. The issue is the opposite — we need more transparency from the companies that offer them and more awareness from developers who build on them.
A free tier makes sense only when it is treated as an integral part of the product — one that cannot be abandoned without a clear and transparent exit strategy, communicated during registration. Users deserve a painless path to migrate services and the information needed to make informed decisions about the platforms they depend on.
Free tiers should attract users, not trap them. There is an abysmal difference between replacing $5 per month with $39 per month. Ending a service is difficult; charging an excessive price on top of that compounds the insult to loyal users. We can do better, and there are plenty of less-cruel alternatives for marketing a product.
