Clerk Pricing Explained

How much does Clerk cost — and why is almost every number you'll find for it wrong?

Clerk's free Hobby plan now covers 50,000 monthly retained users, not the "10,000 users" still quoted across blog posts and competitor comparison pages. Paid plans start at $25/mo for Pro and $300/mo for Business, with Enterprise a custom annual quote. The number that trips people up most is cost at scale: at 100,000 retained users you'll pay about $1,025/mo on Pro — not the roughly $1,825 stale third-party math still produces. The figures disagree because Clerk bills monthly retained users (MRU), a deliberately narrower unit than the monthly active users (MAU) that Auth0, Supabase, and most others meter, and because Clerk replaced its long-running 10,000-user free tier with the 50,000-MRU tier on February 5, 2026 (Clerk pricing; Clerk plans changelog).

This guide gives you a defensible, re-derivable number for Clerk and its main alternatives. It explains the MRU model in plain English, breaks down every plan, pre-calculates costs at 1,000 / 10,000 / 50,000 / 100,000 users for Clerk, Auth0, WorkOS, and Supabase, and is candid about where Clerk is not the cheapest option. Every price is cited inline to a primary source and was verified in June 2026; authentication pricing changes often, so re-check each vendor's page before you budget.

TL;DR: Clerk pricing at a glance

The headline for most teams: Clerk is free to 50,000 retained users, $25/mo for production features on Pro, and predictable from there because the overage rate is published all the way to 1,000,000+ users. The table below puts Clerk next to Auth0, WorkOS, and Supabase at four user milestones. One caveat — these vendors bill different units (MRU vs MAU) and bundle different things, so the raw monthly number is where the comparison starts, not where it ends.

All figures are monthly USD list prices, verified June 2026 against each vendor's pricing page; Clerk and Supabase rows assume the paid production tier, with free tiers in parentheses. ¹ Auth0 prices in discrete MAU tiers (no per-MAU rate): B2C Essentials self-serves up to 50,000 MAU at $3,500/mo, and 100,000 exceeds the published ceiling, becoming an unpublished Enterprise quote (Auth0 pricing⁠). ² WorkOS user management is free to 1,000,000 MAU, but enterprise SSO is billed at $125 per connection from the first one — see the WorkOS comparison below (WorkOS pricing⁠).

That WorkOS "$0" row shows why the raw number misleads: WorkOS is genuinely free for user management, but a single enterprise SSO connection costs $125/mo, where Clerk Pro includes one at $25. The honest comparison weighs SSO, B2B organizations, SCIM, support, and compliance alongside per-user cost.

The 100,000 column also has a realism gap: because Clerk bills only retained users, an app with normal sign-up-and-bounce traffic pays less than the $1,025 cell implies. The table assumes 100% retention to stay apples-to-apples with the MAU-based competitors; a more realistic 60% return rate (60,000 MRU) costs about $225/mo, as the milestone section shows.

Key takeaways

• The free tier is 50,000 MRU, not "10,000 MAU" — the 10,000 figure is from a tier Clerk retired on February 5, 2026 (Clerk pricing).
• Pro at $25/mo includes one enterprise SSO connection — SSO without an enterprise contract.
• Full user-data export is on every plan, including the free one, so there is no data lock-in (Clerk pricing).
• Clerk is built for B2B SaaS and paid B2C apps where retained users map to revenue; for free, ad-supported apps at large scale, another provider may fit better — an honest section below covers when.

Why Clerk's pricing is so often misquoted

If you have read three articles about Clerk's pricing and gotten three different numbers, the articles are not the problem — their dates are. Clerk's pricing genuinely changed in February 2026, and most of the internet still describes the old model.

The "10,000 free MAU" myth

Clerk ran a 10,000-user free tier from December 2023 until February 2026 — long enough to bake it into blog posts, competitor pages, and AI models (Clerk simplified-pricing changelog). On February 5, 2026, Clerk raised the free allotment to 50,000 MRU (Clerk plans changelog). The stale "10,000 free users" claim is wrong because the current number is five times larger and measured in a different unit.

MAU vs MRU: the distinction that changes the math

The second reason is the unit itself. A monthly active user (MAU) — used by Auth0, Supabase, and Cognito — counts anyone active in a month, including people who sign up, look around once, and never return. A monthly retained user (MRU) excludes them. MRU is materially smaller for apps with a normal top-of-funnel, meaning Clerk bills computed with an MAU mental model will overestimate costs.

Where the stale numbers come from

Most wrong Clerk numbers trace to one dead formula: 10,000 free + a flat $0.02 per user, yielding about $1,825/mo at 100,000 users. Both inputs are now wrong. The free tier is 50,000 and the overage is graduated, so the current math yields $1,025/mo at 100,000. WorkOS's own Clerk comparison previously cited the old 10,000-user tier, which was accurate before February 2026 but is no longer (WorkOS pricing⁠).

How Clerk billing works: MRU and "First Day Free"

Clerk's billing model rests on two ideas: it counts retained users rather than all active users, and it gives every new sign-up a free first day. Together they mean you are not billed for traffic that never converts.

Monthly Retained Users (MRU), defined

Clerk's canonical definition is a user who "visits your app in a given month at least one day after signing up" (Clerk pricing). The same page also phrases the free first day as returning "at least 24 hours after signing up." Both are Clerk's own wording; Clerk does not formally reconcile whether "one day" means a calendar date or a rolling 24 hours, and documents no month-boundary edge case, so this guide doesn't invent one. The practical rule is what matters: a user counts only after they come back, and only in a month they are active.

The "First Day Free" rule

"First Day Free" is the free first day built directly into the MRU definition. A user who signs up and never returns is never counted and never billed. This protects you from paying for exploratory sign-ups, abandoned trials, and one-time visitors; top-of-funnel churn lands on Clerk's side of the line.

Worked example: from 1,000 sign-ups to 100,000

Suppose your app gets 1,000 new sign-ups in a month: 300 never return after day one, and separately 200 users from earlier months are inactive this month.

• Clerk (MRU): the 300 first-day bouncers and the 200 inactive prior users are not counted. Clerk counts the ~700 who returned and were active — comfortably inside the free tier, so $0 on Hobby.
• An MAU model (Auth0 and most others): every sign-up logged in at least once, so all 1,000 count as active, including the 300 who never came back.

The mechanic only turns into money at scale. Suppose 100,000 people sign up over time and, at a 60% return rate, 60,000 are retained in a given month (an illustrative rate, not a Clerk-published figure — use your own data). On Clerk Pro:

A pure MAU model would meter all 100,000 — charging for the 40,000 who signed up but never returned. That gap, free under Clerk's model, is the entire point of MRU.

What counts as a retained user (and what doesn't)

Clerk plans, explained (Hobby, Pro, Business, Enterprise)

Clerk publishes four plans — Hobby, Pro, Business, and Enterprise — and the right one is usually obvious from where you are in your product's life. Pricing below is from Clerk's pricing page (Clerk pricing).

Hobby (free) — best for indie projects, prototypes, and early MVPs

Hobby is $0 and includes 50,000 MRU. It's a real free tier scoped for early-stage work: up to 3 social connections, a 7-day session, and 1-day log retention, but no multi-factor authentication or production passkeys. Exceed 50,000 MRU and you get a one-month grace period to upgrade.

Pro — best for production B2C and small B2B apps

Pro is $25/mo ($20 annually) for production apps. It includes the 50,000-MRU allotment, unlimited social connections, MFA, passkeys, custom sessions, metered SMS, white-labeling, and one enterprise SSO connection. Add-ons include satellite domains ($10/mo) and B2B Authentication ($100/mo) which unlocks unlimited org members, verified domains, and custom roles.

Business — best for scaling teams

Business is $300/mo ($250 annually). It adds 10 dashboard seats, priority support, 30-day log retention, and access to SOC 2/HIPAA compliance artifacts for growing teams.

Enterprise — best for large organizations

Enterprise is custom, annual-only, and unpublished. It adds a 99.99% uptime SLA, HIPAA compliance with a BAA, and volume discounting. This guide doesn't estimate an Enterprise number since it's completely bespoke.

What's included on every plan

Some capabilities are on every plan, including the free one — the items behind the "real product on the free tier, no lock-in" claim:

What Clerk costs at 1,000, 10,000, 50,000, and 100,000 users

Assumptions behind these numbers

Each row assumes: every user is billable (100% MRU, which overstates a real Clerk bill once non-returning sign-ups drop out); a production-grade plan (Pro, because MFA and passkeys are gated above Hobby); and no add-ons. All figures are from Clerk's pricing page (Clerk pricing).

Clerk cost at each milestone

The overage above 50,000 is graduated, not flat: $0.02/user from 50,001–100,000, $0.018 from 100,001–1,000,000, $0.015 from 1,000,001–10,000,000, and $0.012 beyond (Clerk pricing). Those bands are why the real 100,000-user cost is about $1,025/mo, not the ~$1,825 the old flat-rate formula produces.

How to estimate your own monthly cost

Walk the chain sign-ups → retained users (MRU) → plan tier → add-ons. For a Pro app under 100,000 retained users, the formula is $25 + max(0, MRU − 50,000) × $0.02. Add $100/mo for the B2B Authentication add-on (organization-scoped SSO) and $10/mo per satellite domain. Above 100,000 MRU, use the next graduated band for the portion over the threshold.

Clerk vs Auth0, WorkOS, and Supabase

All four solve authentication, but they bill different units and bundle different things, so "cheaper" depends on what you are building. Every figure below was verified live in June 2026; all four restructured pricing between 2024 and 2026, so re-derive from each vendor's page before you budget.

Clerk vs Auth0 pricing

Auth0 bills MAU in discrete tier buckets, not a per-MAU rate. Its pricing states it's "available only at the listed MAU tiers," and between-tier usage is billed at the next tier up (Auth0 pricing⁠). The widely cited "$0.07/MAU" figure is third-party back-math from the 1,000-MAU tier, not an official rate.

B2C Essentials is $70 at 1,000 MAU, $700 at 10,000, and $3,500 at 50,000 (its self-serve ceiling). Compared to Clerk Pro's flat $25 up to 50,000 MRU, Auth0 is significantly more expensive at scale. Auth0 counts every active user, and its fixed tiers step up steeply. For B2B, Auth0 B2B Essentials is $300 at 1,000 MAU and $2,100 at 10,000 MAU with 3 enterprise connections, versus Clerk Pro at $25 with one (Auth0 pricing⁠).

Auth0's strengths include broad compliance, mature authorization, and free inbound SCIM/SSO on every plan, including one enterprise connection on Free (Auth0 B2B plans upgraded⁠). However, Auth0's free tier has no MFA factors (only passkeys), forcing production apps requiring MFA to upgrade. Overall, Auth0 costs more per user and steps up steeply, but it's a deep platform.

Clerk vs WorkOS pricing

WorkOS AuthKit user management is free to 1,000,000 MAU, then $2,500/mo per additional million (WorkOS pricing⁠) — cheaper than Clerk for pure volume. But it meters enterprise SSO per connection: $125/mo for connections 1–15, stepping down at higher volumes. The choice depends entirely on whether you need enterprise SSO.

• No SSO, lots of users: WorkOS's free-to-1M tier wins on raw cost.
• Any B2B app needing even one SSO connection: Clerk Pro includes one at $25/mo, versus $125/mo on WorkOS — about 5× cheaper for the SSO-inclusive case.

WorkOS's startup program and "Annual Credits" prepay plan don't fundamentally change this math on the published per-connection rate.

Clerk vs Supabase Auth pricing

Supabase bundles auth into a backend platform (Postgres, Storage). Its Free tier covers 50,000 MAU (idle projects auto-pause), and Pro ($25/mo) includes 100,000 MAU before a $0.00325/MAU overage (Supabase pricing⁠). It's hard to beat for cost-sensitive B2C apps already using Postgres.

Supabase SAML SSO is Pro-and-up only, billed at $0.015 per SSO MAU after 50 included, and has no SCIM (Supabase pricing⁠). Claims that "Supabase Free includes SAML" are false. Supabase fits cost-sensitive, bundled-stack teams, while Clerk fits those needing B2B organizations, SSO, and directory sync.

How to read these comparisons

Three caveats apply: verify figures before committing; comparisons assume the same user cohort; and raw per-user cost isn't total cost of ownership. SSO inclusion, SCIM, B2B orgs, and compliance-report access often outweigh the per-user line. Seemingly "free" providers can cost more once enterprise features are added.

SSO and enterprise features without enterprise pricing

SSO is included on the Pro plan

Clerk includes one enterprise SSO connection — SAML, OpenID Connect (OIDC), or EASIE (Google/Microsoft) — on the $25/mo Pro plan, with additional connections metered by volume: $75 each for connections 2–15, $60 for 16–100, $30 for 101–500, and $15 beyond 500 (Clerk pricing). That matters because SSO is where competitors most often install a toll booth: WorkOS charges $125 per connection from the very first one, while Supabase, Firebase, and AWS Cognito meter federated SSO at about $0.015 per MAU. The public sso.tax catalog documents the pattern — roughly 150 vendors charging a steep premium to turn on a security baseline (sso.tax / sso-wall-of-shame⁠). Those markups are mostly margin: delivering SAML costs only integration and maintenance labor, the commodity rate near $0.015/MAU.

The dollar gap is easiest to see at a fixed scale. Here is the monthly cost of enterprise SSO for a B2B app with about 10,000 users, on each provider's production tier:

These rows assume a production tier with MFA — the apples-to-apples bar. One fair footnote: Auth0's free tier now includes one enterprise connection (and free inbound SCIM) up to 25,000 MAU, but without MFA factors — so an app that genuinely needs no MFA can run a single Auth0 connection for $0, while a production app that needs MFA lands on the paid tiers above (Auth0 pricing⁠; Clerk pricing; WorkOS pricing⁠).

Other commonly overlooked inclusions

Surface-level comparisons miss value bundled into Clerk's plans rather than sold separately:

• SCIM / directory sync is included with an enterprise connection at no extra charge — at several competitors it is a separate per-connection fee or an enterprise-tier gate.
• Full user-data export on every plan, including Hobby (the no-lock-in guarantee, covered next).
• Published pricing all the way to 1,000,000+ users, so you can model your bill from the public page rather than a sales call.
• 5 free user impersonations per month on paid plans, for support and debugging.

Inclusions evolve, so confirm the current specifics against the pricing page when you plan (Clerk pricing).

When Clerk is NOT the best fit (an honest assessment)

Clerk is built for B2B SaaS and paid B2C apps, and it is not the right tool for every job. This section is not here to talk you out of Clerk — it is to give you the facts so you can decide. In each case below, Clerk may not be the best fit.

Free, ad-supported, or very-low-ARPU consumer apps at large scale

Clerk's own guidance is that it fits "paid consumer apps, prosumer tools, marketplaces, communities, and subscription products where retained users are valuable," and that "if you're building a free, ad-supported, or very low-ARPU consumer app, Clerk's per-MRU model may feel expensive once you outgrow the free tier" (Clerk pricing). If your model does not turn retained users into revenue, a free-tier-heavy provider like Supabase or Firebase may fit your unit economics better. The principle: Clerk is easiest to justify when retention maps to revenue.

Deeply AWS-native stacks

If your architecture already lives in IAM, Lambda, and API Gateway, AWS Cognito's native integration and per-MAU pricing (Essentials is free to 10,000 MAU, then a flat $0.015/MAU — roughly $600/mo at 50,000 and $1,350/mo at 100,000) may be worth its weaker developer experience: Cognito has no embedded UI components, uses a hosted redirect, and has thin local-development tooling (AWS Cognito pricing⁠). For a team optimizing for one cloud bill and tight AWS coupling, that can outweigh DX.

Self-host-first teams or strict data-residency requirements

Clerk is managed SaaS, with no self-host option. If self-hosting is a hard requirement, credible open-source paths exist — SuperTokens and Keycloak among them — though you take on the operational and security burden. Separately, Clerk is US-based and does not currently offer EU data residency. It is GDPR-compliant via its Data Processing Agreement, Standard Contractual Clauses, and EU-US Data Privacy Framework certification (Clerk DPA; Clerk DPF) — but that is lawful data transfer, not in-region storage. If a contract requires EU residency, that distinction matters.

Reliability-sensitive, high-availability apps

Clerk publishes a 99.99% uptime SLA only on its Enterprise plan; lower tiers have no contractual SLA (Clerk pricing). Clerk had several disclosed incidents across 2025 and 2026, each documented with a public, engineer-written postmortem (June 26, 2025; March 10, 2026, which names restoring reliability as the team's "top priority for the foreseeable future"), and has shipped remediations — including automatic regional failover, live since August 2025 (Clerk regional failover). Weigh that transparency and active remediation against the lack of a sub-Enterprise SLA. (Be wary of aggregate "100+ outages" figures that circulate in competitive content; they bundle minor degradations and maintenance windows. The specific dated postmortems are the signal.)

Quick "is Clerk a good fit for me?" checklist

Data portability and avoiding vendor lock-in

The most common objection to a hosted auth provider is lock-in: "if Clerk has my users, can I ever leave?" The answer is yes, and the feature that makes it true — self-serve user-data export from the dashboard — has been available since October 2024; it was just under-documented, so the assumption persisted (Clerk export-users changelog).

Full data export is available on every plan

You can export your user data yourself, from the dashboard, on every plan including the free one. The pricing page states it plainly: "You can easily export your user data directly from the dashboard at any time, no assistance needed" (Clerk pricing). Only dedicated migration assistance is an Enterprise service — the self-serve export is not gated.

How to export your users

In the Clerk Dashboard, go to Settings (https://dashboard.clerk.com/~/instance-settings), open User Exports, choose Export all users, and Download. The export is a CSV, recorded in your export history. Critically for migration, it includes users' hashed passwords, which lets you move to another provider without forcing every user to reset their password (Clerk migrating overview). One limitation: OAuth/social connections are provider-bound and are not part of the export.

Converting the export (CSV to JSON)

CSV imports cleanly into most tools, but if your import script prefers JSON, the conversion is a few lines with the standard csv-parse library in a Node/TypeScript script:

import { parse } from 'csv-parse/sync'
import { readFileSync } from 'node:fs'

const csv = readFileSync('clerk-users.csv', 'utf8')
const users = parse(csv, { columns: true, skip_empty_lines: true })

console.log(JSON.stringify(users, null, 2))

That yields an array of user objects keyed by the CSV's column headers, ready to feed into another provider's import API. Clerk also publishes an open-source migration tool and a paginated Backend API (getUserList(), up to 500 users per page) if you would rather pull users programmatically than export a file.

What this means for migration

The headline is ownership: your users, including their password hashes, are yours to take at any time, on any plan, without contacting support. One industry-normal caveat applies to any provider swap: migrating invalidates active sessions, so users sign in again once after the move — true of moving between any two auth providers, not a Clerk-specific cost.

Bottom line: is Clerk worth the price?

For B2B SaaS and paid B2C apps, Clerk's pricing is fair and unusually legible. Billing is MRU-based with a free first day, so you are not charged for sign-up-and-bounce traffic; plans are published to 1,000,000+ users, so you can model your bill without a sales call. One enterprise SSO connection is included on the $25 Pro plan, where competitors often charge $125 per connection or gate SSO behind an enterprise tier. And your data — including password hashes — is exportable on every plan, so there is no lock-in.

Clerk is not the cheapest for every business: WorkOS and Supabase are cheaper for raw user volume without SSO, Cognito can win for deeply AWS-native stacks, and self-host or strict EU-residency requirements point elsewhere. But for the common case — a product whose users map to revenue and whose roadmap includes B2B organizations and enterprise SSO — Clerk's model is straightforward to budget and hard to beat on SSO-inclusive cost. Start on Hobby, move to Pro when you need production features, and check the current numbers on the Clerk pricing page before you commit.