Questions

Because its pricing page publishes $35/mo and $240/mo at 500 monthly active users and then hands the rest to a slider widget rather than a table. This catalogue records what the vendor publishes as text, and records the absence honestly instead of guessing.

Also answered on /clerk-vs-auth0

At the bottom, Auth0: its free tier runs to 25,000 monthly active users where Clerk's is a hard cap at 50,000 MRU with Clerk branding on the components, and Clerk's first paid tier is $25/mo against Auth0's $35 / month. Above that the honest answer is that only one of them can be worked out. Clerk publishes $0.02 per monthly active user stepping to $0.015; Auth0's page carries no per-user rate at all above 500 monthly active users.

Also answered on /clerk-vs-auth0

WorkOS, on the published terms. It prices SAML by the connection: $125 / ea, falling to $65 each above 50 connections: while AuthKit itself is free to 1 million users, so the bill tracks the number of enterprise customers rather than the number of end users. Auth0 includes one enterprise connection on its free tier and Clerk sells Enhanced B2B SaaS as a $100/mo add-on on top of Pro.

Also answered on /clerk-vs-auth0

It varies by provider, which is a real problem when comparing them. Most count a user who authenticated at least once in the billing month. Read each definition before trusting these figures at the margin: a provider that counts token refreshes bills differently from one counting logins.

Also answered on /clerk-vs-auth0

Yes, and it is a common pairing: Better Auth writes its tables into whichever Postgres you point it at, so you get your own user table with no auth bill on top. The recommender suggests exactly this combination when you ask to keep users in your own database.

Also answered on /neon-vs-supabase

On an idle or bursty workload, usually: Neon's compute suspends when nothing is talking to it and there is no base fee on the Launch plan, so the floor is close to zero. On a database that is busy all month, Supabase's flat $ 25 / month with $10 of included compute credit is often the cheaper shape. Neither is cheaper in general, which is why this page prices both rather than naming a winner.

Also answered on /neon-vs-supabase

Yes. Both are Postgres you can point any Postgres client at, run migrations against, and take somewhere else: which is the thing that separates this pair from a document store. Supabase is open source and self-hostable on top of that, which makes the option to leave credible rather than theoretical.

Also answered on /neon-vs-supabase

It means the compute behind the database suspends when there is no traffic and resumes on the next connection, so you are billed for CU-hours you actually used rather than for a server sitting idle. The cost is a cold start on the first query after a quiet period. On a staging or preview database that is free money; on a low-traffic production API it is a latency decision you should make deliberately.

Also answered on /neon-vs-supabase

At small scale both are free. At 100,000 users the honest answer is that nobody can tell you, because Firebase does not publish a per-user rate above its free quota. Supabase can be worked out to the cent from published figures, which is a different kind of advantage than being cheaper.

Also answered on /supabase-vs-firebase

Not usefully: Supabase Auth writes into the Postgres instance it comes with. If you want auth separate from your database, that is the case for Clerk or Better Auth instead.

Also answered on /supabase-vs-firebase

On the authentication line, its own pricing page says "No-cost up to 50K MAUs: Then Google Cloud pricing" and prints no figure. Storage and transfer do carry rates: "Then $0.026/GB" stored and "Then $0.12/GB" downloaded: so parts of a Firebase bill are predictable and the user-count part is not. The Spark plan's no-cost quotas are also enforced daily and reset at midnight, which means a busy day stops the app rather than billing you.

Also answered on /supabase-vs-firebase

Supabase, and not marginally. It is Postgres, so a dump and restore moves the data to any other Postgres host, and Supabase itself is open source and self-hostable. Firestore is a document store with a query model that forbids most joins, so leaving it means rewriting the data access rather than moving the data.

Also answered on /supabase-vs-firebase

Some do: Stripe for domestic sales, a merchant of record internationally. It works, and it doubles your reconciliation, webhook handling and subscription logic. Worth it only when the saved percentage clearly exceeds the engineering.

Also answered on /stripe-vs-paddle

Yes, and knowingly so: 5% + 50¢ per Checkout transaction against Stripe's 2.9% + 30¢ for domestic cards. The gap narrows once you add Stripe's published surcharges: "+ 1.5% for international cards" and "+ 1% if currency conversion is required", and it narrows further once you count what you would pay an accountant to file VAT in the EU and UK. It does not close.

Also answered on /stripe-vs-paddle

It becomes the legal seller of your product, so the customer buys from Paddle rather than from you. That means Paddle collects, files and remits VAT, GST and sales tax everywhere it is owed, and carries the liability if that is wrong. Stripe processes the payment and leaves all of that with you.

Also answered on /stripe-vs-paddle

Yes, from the first euro: there is no small-seller threshold for non-EU businesses selling digital services into the EU. That obligation is the entire reason the merchant-of-record category exists. Whether anyone comes looking is a different question from whether it is owed.

Also answered on /stripe-vs-paddle

It is a real consideration. A merchant of record holds your revenue between the customer paying and you being paid, so its solvency and payout reliability matter more than a percentage point. Weigh the saving against how long the company has been doing this.

Also answered on /polar-vs-lemon-squeezy

About $1,379 a month in sales, and that figure is Polar's own: its fees page publishes the break-even for each paid tier: "~$1,379 /mo in sales" for Pro, "~$5,634 /mo" for Growth, "~$19,048 /mo" for Scale. Below the first threshold the two are the same 5% + 50¢ and the monthly fee is a straight loss.

Also answered on /polar-vs-lemon-squeezy

At the headline rate, neither: both are 5% + 50¢ with no monthly fee. The difference is structural: Polar sells tiers that trade a fixed monthly fee for a lower variable rate, and Lemon Squeezy does not sell tiers at all. So Lemon Squeezy is cheaper for anyone under about $1,379/mo in sales, and Polar is cheaper for everyone above it.

Also answered on /polar-vs-lemon-squeezy

Yes: that is what makes both of them a merchant of record rather than a payment processor. Each becomes the legal seller, collects and remits the tax, and carries the liability. That is the service the 5% is buying; a plain processor at 2.9% leaves the filing with you.

Also answered on /polar-vs-lemon-squeezy

Usually yes, and the friction is in the edges rather than the app: image optimisation, middleware behaviour, ISR revalidation and cron jobs all differ. Budget a day, not an afternoon.

Also answered on /vercel-vs-netlify

It depends entirely on how many people are on the team. Netlify's Personal plan is $9 /month for one member and Pro is $20 /month for unlimited members; Vercel's Pro is $20 /mo. per seat. One developer, the two are close. Eight developers on a modest site, Netlify is roughly a quarter of the price. On heavy metered usage the comparison reverses and has to be done in the calculator, because Netlify meters in credits and Vercel meters in gigabytes and invocations.

Also answered on /vercel-vs-netlify

Because Hobby is non-commercial only. A plan that forbids the thing you are doing is excluded before price is considered: recommending it to someone charging money for their app would be recommending a terms-of-service violation. Netlify's "$0 forever" tier carries no such restriction.

Also answered on /vercel-vs-netlify

Fast Data Transfer is 100 GB included on Hobby, and on Pro it is covered by flat rate CDN rather than billed per GB: "CDN requests, fast data transfer, and other resources are no longer billed directly." The metered rates that remain on Pro include function invocations starting at $0.60 per 1M, Fluid active CPU starting at $0.128 per hour and Fluid provisioned memory starting at $0.0106 per GB-hour. Netlify's overage is denominated in credits instead: "$5 / 500 additional credits" on Personal and $10 / 1,500 on Pro.

Also answered on /vercel-vs-netlify

Because it is non-commercial only. A plan that forbids the thing you are doing is excluded before price is considered: recommending it to someone charging money for their app would be recommending a terms-of-service violation.

Also answered on /vercel-alternatives

On the invoice, usually yes. The comparison this page cannot make for you is the cost of your own time doing the parts these platforms do: deploys, TLS, rollbacks, scaling and the 3am page. That is a real cost that does not appear on any pricing page.

Also answered on /vercel-alternatives

Cloudflare Workers at $5.00 a month for 10 million requests, if your app fits the Workers model. DigitalOcean App Platform is "Paid Tier Starting at $5 /month" per container. Netlify's Personal plan is $9 /month for one member. The cheapest that is also free is Netlify's "$0 forever" tier or Cloudflare's 100,000 requests a day: Vercel's own free tier is disqualified for commercial use and Railway publishes no free plan.

Also answered on /vercel-alternatives

Netlify, if you are leaving on price or team size: it runs Next.js and the migration friction is in image optimisation, middleware and ISR rather than the app code. Railway or Render if the reason you are leaving is that you need a process that stays running, which is the thing a serverless platform is structurally bad at. Cloudflare Workers is the best answer for bandwidth and the worst for a large Node application, because the runtime is not Node.

Also answered on /vercel-alternatives

No. This catalogue only carries products with a public rate card, and Auth.js does not publish one. Better Auth is in the table as the reference point for the self-hosted option. Its framework is free and open source, and Better Auth Inc. publishes a rate card for its managed infrastructure at better-auth.com/pricing, which this catalogue reads nightly.

Also answered on /best-auth-for-nextjs

It varies by provider, which is a real problem when comparing them. Most count a user who authenticated at least once in the billing month. Read the definition before trusting any of these numbers at the margin: a provider that counts token refreshes will bill differently than one counting logins.

Also answered on /best-auth-for-nextjs

At any realistic small-app scale, all of them are free: the smallest allowance in this table is 10,000 monthly active users and WorkOS gives away a million. The first plan you actually pay for is Clerk or Kinde at $25 a month, then Auth0's Essentials at $35 / month, then PropelAuth's Growth at $150 /month. Better Auth run as a library is $0 forever, because the users live in a database you already pay for.

Also answered on /best-auth-for-nextjs

All of them here except Auth0. Clerk publishes "$0.02/mo each" stepping to $0.015, Kinde "$0.0175 per extra MAU", PropelAuth "$0.05 per user", WorkOS "$2,500 / mo" per additional million, Stytch "$125/connection", Better Auth "then $0.0001 per event". Auth0's page prints $35 / month and $240 / month at 500 monthly active users and moves the rest into a slider, so there is no rate on the page to multiply out.

Also answered on /best-auth-for-nextjs

Yes, from the first euro: there is no small-seller threshold for non-EU businesses selling digital services into the EU. This is the entire reason the merchant-of-record category exists. Whether anyone comes looking is a different question from whether it is owed.

Also answered on /merchant-of-record-for-saas

You can, and the subscription migration is the painful part rather than the integration. Moving live subscriptions between processors means re-collecting payment details or arranging a card-network migration. Choosing on rate alone and planning to switch at scale usually costs more than it saves.

Also answered on /merchant-of-record-for-saas

Creem, on published rates: 3.9% + $0.40 per successful transaction with no monthly fee, and its comparison table claims "No international card fees", which matters more than the headline gap on a customer base outside the United States. Dodo Payments is 4% + 40¢ but adds "+1.5%" for international cards and $30 per dispute. Polar and Lemon Squeezy both start at 5% + 50¢; Polar's paid tiers take that down to 3.4% + 30¢ at $400 /mo, which pays off above about $19,048 a month in sales.

Also answered on /merchant-of-record-for-saas

It depends on where your customers are, and the arithmetic is genuinely close. Two points of $10,000 a month is $200: roughly what an accountant charges to file EU VAT quarterly, and less than the cost of getting it wrong. Below that revenue the merchant of record is usually cheaper than the alternative; well above it, and with customers concentrated in one or two jurisdictions, doing it yourself on Stripe starts to win.

Also answered on /merchant-of-record-for-saas