How Much Does It Cost to Build an App LikeCalm in 2026?

A lean first version of an app like Calm comes to about 890 to 1,470 hours of senior work in my estimate, which at $150 to $225 an hour is $133,500 to $330,750. I'll show every one of those hours below, and you can check the addition yourself.

A fuller launch, with Apple Health sync, a watch companion, personalization, several languages, a web player and an employer plan, lands at about 1,540 to 2,620 hours, or $231,000 to $589,500.

Those are my scenario estimates. They are not quotes, and they are not measurements of what Calm spent. Calm is a trademark of its owner, and nothing here implies any relationship with the company.

Here's the thing that surprises people about this category. The software is the smaller problem. A meditation app is, mechanically, a very good audio player wrapped in a subscription. What makes Calm Calm is hundreds of hours of recorded content, and that content is a separate budget that no development quote covers.

Consider a founder who wants "Calm, but for nurses," or for new parents, or for athletes before a fight. The first thing that founder usually reads is an agency page saying the whole thing costs somewhere between $30,000 and $250,000. That range is more than eight times wide. It tells you nothing you can plan around.

So here is the method instead. List the features the real product has, split what a v1 needs from what Calm built over years, price each feature in hours, then read the running costs straight off the vendors' own price pages. Every number you see is either cited, one of our own published rates, or labelled as my arithmetic.

The published agency ranges for a Calm clone cannot be reproduced from anything on the pages that print them, so I don't repeat them as fact. Here are four, cited as examples of the claim rather than as evidence.

VLinkmanages two answers on one page. It says a meditation app like Calm costs around $30,000 to $35,000 on average, and elsewhere that a complete development cycle for Calm or Headspace runs $57,000 to $65,000. It does show one piece of arithmetic, 800 hours at $25 an hour for $20,000, and lists US developer rates of $80 to $150. But 800 hours at $80 is already $64,000, and at $150 it is $120,000. The inputs and the headline don't describe the same project.

Chop Dawg prints $40,000 to $85,000 for an MVP and $120,000 to $250,000 plus for a full clone. It breaks the app into modules, which is the right instinct, but gives no hours and no rate, so there is nothing to multiply.

PerfectionGeeks gives $30,000 to $250,000, again without hours or a rate. It opens with two mental health percentages that carry no source at all.

RaftLabsis the most careful of the four and still prints bands ($40,000 to $75,000 for an MVP, up to $220,000 for a scale build) with no hour count. It also repeats a Calm revenue figure and a valuation credited to a third-party app analytics publisher, plus a market-size forecast from a research firm. Calm doesn't publish audited revenue, for reasons I'll get to in a minute, so those numbers are estimates about a private company. I don't print them.

The same goes for the numbers these pages open with: meditation app market sizes, growth rates, download counts, "X% of adults are stressed." I looked for a primary source behind each one I found and could not fetch one that supports it. None of them appear in this article. If a vendor puts one in a pitch deck for you, ask for the underlying report.

This is the same refusal I made across this series. The Spotify cost breakdownruns the method on the other big audio product, and a lot of the player work overlaps. What doesn't overlap is licensing: Spotify rents music from labels, Calm mostly commissions its own recordings.

Calm is a private company, so there is no 10-K to read and no audited R&D number to compare against. That matters, because in the other articles in this series I anchor the estimate against a public company's filed engineering spend. Here I can't.

What I can see is this. Calm.com, Inc. appears on SEC EDGAR under CIK 0001721834, and its filing index lists Form D notices, which private companies file when they raise money under an exemption, plus a few ownership forms. No annual report. No registration statement.

The most recent Form D, filed December 23, 2020, reports a $40,000,000 offering with $39,499,920 sold and a first sale date of December 10, 2020. That tells you the company raised serious money. It tells you nothing about how much of it went to engineering, and I won't pretend otherwise.

So I'll build the feature inventory from what Calm shows the public. Its App Store listing describes guided meditation, Sleep Stories, breathwork, music and soundscapes, with named narrators. It lists a $14.99 monthly plan and a $69.99 annual plan among its in-app purchases, says there are never any ads, runs on iPhone, iPad, Mac, Apple TV, Apple Watch and Apple Vision, syncs mindful and sleep sessions to Apple Health, and ships in English plus nine other languages.

Calm's Help Center explains offline listening: you tap a download icon on a program or in the player, and your saved sessions live under Profile, Library, My Downloads. That one feature, it turns out, is a real chunk of the build.

Here is the inventory split the way I'd split it for a client.

The second list is where the years went. The first list is what you're actually pricing. Think of it like opening one good yoga studio versus buying a national chain. Same category, completely different balance sheet, and nobody sizes the studio from the chain's investor deck.

The lean v1 adds up to 890 to 1,470 hours in my estimate. That assumes iOS and Android from one cross-platform codebase, a web-based admin, a managed backend, and a senior team. Every row is my scenario estimate, not a quote.

Let me check the sum out loud, because a table that doesn't add up is the same sin as the agency ranges. Low column: 60 + 120 + 120 + 80 + 60 + 40 + 100 + 30 + 40 + 120 + 120 = 890. High column: 100 + 200 + 200 + 140 + 100 + 70 + 160 + 50 + 70 + 200 + 180 = 1,470.

Now the rows people underestimate, and why.

The player. Pressing play is easy. Keeping audio going when the phone locks, when a call comes in, when AirPods disconnect, when the user sets a 30 minute sleep timer and falls asleep at minute 12, and resuming exactly where they left off the next night: that is where the 120 to 200 hours go. Sleep content makes this harder, not easier, because the whole point is that nobody is watching the screen.

Offline downloads.A download feature has to decide where files live, how much storage they may take, what happens when the phone runs low, and what happens when a subscription lapses and the downloaded premium content should stop playing. That last one is the part founders forget. You're building a small rights system, not a save button.

The admin. I put 120 to 200 hours on content tooling because a meditation app lives or dies on publishing new sessions every week. If adding a session requires an engineer, the content team stalls, and a stalled library is how subscribers quietly churn. We built an admin dashboard alongside the iOS and Android apps and a NestJS backend for LIVV Audio, a Bluetooth audio platform, and the pattern holds: the admin is not an afterthought, it is how the business runs day to day.

The subscription. 100 to 160 hours covers the store purchase flows on both platforms, the free trial, a paywall, entitlement checks on the server, and the webhooks that tell your backend when someone renews, cancels, refunds or gets a billing retry. The in-app purchase implementation guide goes through that plumbing in detail. It is also where the subscription laws below turn into code.

Why cross-platform? Because the product is a player and a list, not a game engine. A single React Native codebase with native audio modules covers both stores for a v1, and that is what the hours above assume. If you insist on two fully native apps, add roughly the mobile share again.

The multiplication, so you can check it. 890 x $150 = $133,500. 1,470 x $150 = $220,500. 890 x $225 = $200,250. 1,470 x $225 = $330,750. The full lean band is the lowest of those to the highest: $133,500 to $330,750.

A fuller launch adds about 650 to 1,150 hours, for a total of 1,540 to 2,620. This is the version that starts to look like the Calm people picture: health sync, a watch, recommendations, more than one language, a web player and something to sell to employers.

Checking the added hours. Low: 60 + 100 + 80 + 60 + 100 + 120 + 50 + 80 = 650. High: 100 + 180 + 160 + 100 + 160 + 220 + 90 + 140 = 1,150. Add the lean v1 and you get 890 + 650 = 1,540 and 1,470 + 1,150 = 2,620.

At $150 an hour that is $231,000 to $393,000. At $225 it is $346,500 to $589,500. So the fuller launch band is $231,000 to $589,500.

Two rows deserve a warning. The Apple Watch companion looks small because the watch app is small, but it runs on its own device with its own constraints and its own review. And the employer plan is a second product: seat management, invites, a company admin, invoicing outside the app stores. If B2B is the business model, price it as its own project.

Also notice what the multi-language row does not include. 60 to 100 hours gets you the plumbing for translated menus and per-language content catalogs. It does not get you the content in those languages. That is the next section, and it is the biggest number in the whole project that I can't give you.

If your idea sits closer to live classes and workouts than to sleep and breath, the Peloton cost breakdown in this series prices live streaming and leaderboards, which change the math a lot.

The content library is usually the largest cost of a meditation app, and its price is not publicly disclosed. I'm going to say that plainly rather than invent a per-session figure.

Think about what Calm's listing actually sells: named narrators, original music, sleep stories written for the purpose. Each of those is a writer, a voice, a studio session, an edit and a license, and every one of those fees is negotiated privately. I went looking for a published rate sheet that covers meditation or sleep narration and didn't find one I could verify. Union rate sheets exist for audio commercials, but a commercial is a different kind of work, and reusing its scale for a 30 minute sleep story would be a guess dressed up as a source.

Celebrity narrators are the extreme case. What a famous actor charges to read a bedtime story is not published anywhere I can cite, so I won't estimate it.

What does have a published price is synthetic speech. Google Cloud's Text-to-Speech pricing lists Chirp 3 HD voices at $30 per million characters after the first million free each month, Neural2 at $16 per million and Studio voices at $160 per million.

Let's do the arithmetic for a scenario. Assume a 30 minute session read at about 130 words a minute. That is about 3,900 words. At roughly 5.7 characters a word including spaces, it's about 22,000 characters. A launch library of 200 sessions is then about 4.4 million characters. Take off the free million and Chirp 3 HD costs 3.4 x $30 = $102. Studio voices would cost 3.4 x $160 = $544.

To be clear, that doesn't mean you should launch with robot voices. In this category the voice is the product, and a synthetic narrator may be exactly what your subscribers are paying to escape. But it does show you where the cost is not. The generation is cheap. The writing, the direction and the taste are where the money goes, and you can prototype scripts with synthetic audio before you pay a human to record the ones that test well.

Serving meditation audio is cheap on published prices, often a few dollars a month, because audio files are small and at least one major storage provider doesn't charge for egress. Video is where the bill grows.

Cloudflare R2 lists standard storage at $0.015 per GB-month, write requests at $4.50 per million and read requests at $0.36 per million, with egress to the internet free. Its free tier covers 10 GB-month of storage, 1 million writes and 10 million reads a month. For comparison, Amazon S3 lists standard storage in US East at $0.023 per GB-month for the first 50 TB, and 100 GB a month of free data transfer out aggregated across AWS. CloudFront now also sells flat-rate plans at $0, $15, $200 and $1,000 a month with no overage charges.

Here is a labelled scenario. 50,000 monthly users, each playing 20 sessions a month, averaging 15 minutes. That is 50,000 x 20 x 15 = 15,000,000 minutes of listening.

At 128 kbps, one minute of audio is 128,000 bits a second, divided by 8, times 60 seconds: 960,000 bytes, or about 0.96 MB. So 15,000,000 minutes x 0.96 MB is 14,400,000 MB, about 14.4 TB a month going out. On R2, that egress costs $0.

The library itself is tiny. 500 hours of audio is 30,000 minutes, which is 28.8 GB at 128 kbps. Add a 64 kbps copy for weak connections (14.4 GB) and you store 43.2 GB, billed as 44 GB because R2 rounds up. Minus the 10 free GB, that's 34 x $0.015 = $0.51 a month.

Reads: 50,000 x 20 = 1,000,000 sessions. If each session makes about 20 requests (segments and range requests, downloads included), that is 20 million reads. Minus 10 million free, 10 x $0.36 = $3.60.

Now add short stretching videos, the way Calm's catalog mixes in movement. Cloudflare Stream bills $5 per 1,000 minutes stored and $1 per 1,000 minutes delivered. Say 100 videos of 10 minutes (1,000 minutes stored, $5) and each user watching 2 of them a month: 50,000 x 2 x 10 = 1,000,000 minutes delivered, which is $1,000.

Look at where the money is. The entire audio side is about $4.11 a month. The video side is $1,005. One modest video feature costs roughly 245 times the whole audio operation. If your cost model is tight, keep video out of v1.

Your auth, database, push and analytics vendors add to this, and those vary too much by choice to put in one table. The point stands: for an audio-first product, hosting is a rounding error next to the build and the content.

The app stores are your largest running cost once people pay, and for a small developer both currently take about 15% of a subscription. The exact rate depends on programs and dates, so read the current pages before you model.

Apple's Small Business Program charges 15% for developers with up to $1 million in proceeds in the prior calendar year, and new developers qualify. Separately, Apple's subscription pagesays developers receive 70% of the subscription price in a subscriber's first year and 85% after a year of paid service, and that free trials don't count toward those paid days.

Google Play's service fee page lists 15% for auto-renewing subscriptions, and, for the US, EEA and UK starting June 30, 2026, 10% plus a 5% billing fee for auto-renewing subscriptions under the new fee table. Same 15% total, different structure.

Most teams put a subscription layer between the app and the stores. RevenueCat is free up to $2,500 in monthly tracked revenue and charges 1% of tracked revenue above that.

Worked example, a scenario and not a forecast: 5,000 people on Calm's listed $69.99 annual price, just to use a real public price point.

So a small developer on either store keeps about $297,457.50 of $349,950 before RevenueCat and taxes, and about $293,958 after RevenueCat. Miss the Small Business Program enrollment on Apple and the first-year fee doubles to $104,985 on the same revenue. That's a $52,492.50 mistake, and it costs one form.

Free trials matter here too. Apple allows one introductory offer per subscription group and asks you to show the trial length and the price after the trial in the purchase flow. Hold that thought, because California asks for something similar and more.

The federal click to cancel rule is not in force, but subscription apps still have binding cancellation and consent duties under ROSCA and state automatic renewal laws. This is not legal advice; it is what those obligations force you to build.

Click to cancel: vacated, then restarted. The FTC amended its Negative Option Rule in 2024 with a simple cancellation requirement. On July 8, 2025, the Eighth Circuit vacated the whole amended rule in Custom Communications v. FTC, finding the agency skipped a required preliminary regulatory analysis, days before the July 14, 2025 compliance date (Manatt summary). On March 11, 2026, the FTC announced an advance notice of proposed rulemaking to start again, citing more than 100,000 complaints in five years. As of September 30, 2026, I found no proposed rule text. So the status is: vacated, and at the advance notice stage.

ROSCA: in force. The Restore Online Shoppers' Confidence Act, 15 U.S.C. 8403, requires anyone selling a negative option online to disclose all material terms before taking billing information, get express informed consent before charging, and provide simple mechanisms to stop recurring charges. A free trial that rolls into an annual plan is exactly that kind of offer.

California: in force since July 1, 2025. AB 2863, signed September 24, 2024 as Chapter 515, expands the state's automatic renewal law to cover free-to-pay conversions. It requires express affirmative consent to the renewal terms, keeping proof of that consent for three years or one year after the contract ends (whichever is longer), an annual reminder, cancellation through the same medium the customer used to sign up, and advance notice before a price change.

What that means in code, for a Calm-like app:

• A paywall that states trial length, the price after the trial and the renewal period next to the button, not behind a link
• A consent record stored on your server with a timestamp and the exact terms shown
• An annual reminder email for web subscribers, with a cancel link
• A cancel path on the web if people can sign up on the web
• A price change notice flow before any increase

Store purchases help, because Apple and Google run the cancellation screens for subscriptions bought through them. They don't cover your web checkout. The day you add web sign-ups to dodge store fees, these duties land on your own code. In my estimate that is 40 to 80 extra hours for a compliant web subscription flow, and it is already inside the fuller launch's web checkout row.

A meditation app is usually a wellness product rather than a medical device, but it can still fall under the FTC's health breach rule and its health claims guidance. Three documents set the boundaries.

FDA general wellness. The FDA's final guidance General Wellness: Policy for Low Risk Devices, issued January 2026 under docket FDA-2014-N-1039, describes when the agency will not regulate low-risk products meant only to support a healthy lifestyle. Relaxation, stress management and better sleep sit on the wellness side. Say your app treats insomnia or reduces symptoms of depression and you are walking toward device territory. The cheapest compliance tool here is a copywriter who knows the line.

FTC health claims. The FTC's Health Products Compliance Guidancesays health benefit claims need competent and reliable scientific evidence, generally randomized controlled trials in people, and that testimonials don't substitute. If your marketing says "fall asleep 40% faster," you need a study that shows it. Otherwise, don't say it.

FTC Health Breach Notification Rule. The FTC's rule summary says amendments announced April 26, 2024 took effect July 29, 2024 and make clear the rule applies to most health apps not covered by HIPAA, with penalties of up to $51,744 per violation. Its compliance guide adds that sharing health data without authorization counts as a breach, that notice is due within 60 calendar days, and that apps drawing data from several sources, such as wearables, are likely covered.

That last point is why the Apple Health row matters. The moment your app reads sleep or mindful minutes from Apple Health or Health Connect, you are pulling health data from another source, and your analytics SDKs become a liability. A marketing pixel that sends "user completed an anxiety program" to an ad network is the kind of disclosure that rule is aimed at.

In practice this forces three builds: a data map of every SDK and what it receives, consent gates in front of any third-party analytics that touches health events, and an incident process that can notify users within 60 days. No public figure exists for what that costs, so I put it inside the analytics and consent row (40 to 70 hours) and the health sync row, not as a separate dollar amount.

HIPAA is a different trigger. A consumer meditation app sold directly to the public is usually not a HIPAA business associate. If you sell into a health plan or a provider and handle their patients' data, it changes. Frenchy Digital signs a BAA with healthcare clients, and we don't hold SOC 2, ISO 27001 or HITRUST; if a buyer requires one of those, that is a vendor question to settle first. The wider picture of this category is in the fitness and wellness app development guide.

Year two of a Calm-like app is mostly maintenance hours and new content, since hosting stays small. My scenario is 30 to 50 hours a month of senior maintenance.

Why that much for an app that "just plays audio"? Because both mobile platforms ship major OS versions every year, the store billing libraries change on their own schedules, audio session behavior shifts between OS versions, and every change to the subscription flow has to be retested in both sandboxes. Sleep apps get bug reports at 2am, from people who were asleep when it broke.

The arithmetic: 30 x $150 = $4,500 a month at the low end, 50 x $225 = $11,250 at the high end. Over twelve months, $54,000 to $135,000.

Add hosting from the table above (a few dollars for audio, about $1,000 a month if you keep the video feature in the scenario) and store fees that scale with revenue. Then add new content, which is the line most likely to decide whether year two has subscribers at all, and whose cost, again, depends on privately negotiated fees.

The total cost of ownership guide goes through the full list of recurring lines for any app. For this category the ratio looks unusual: infrastructure is tiny, people and content are almost everything.

A full Calm-like v1 is bigger than any of our starter packages, and I'd rather say that here than on a call. Our published MVP development packages are $15,000 to $25,000, $30,000 to $50,000 and $55,000 to $75,000 plus. The lean scenario in this article starts at $133,500.

Divide it out: $133,500 / $75,000 is about 1.78. Even the bottom of my lean band is nearly 1.8 times the top package. So if someone offers you a Calm clone for a starter package price, something on the lean list isn't in it.

Where a starter package does fit is a narrower first product. One program instead of a library. Twenty sessions instead of two hundred. One platform, a paywall with a trial, and streaming without downloads. That tests whether people will pay for your particular angle, whether it's nurses or fighters or new parents, before you fund a year of content.

Our senior rate is $150 to $225 an hour. We've been building since 2016 (founded in France, a US company since 2019), we're a senior-led, Black-owned team in Los Angeles, and on full payment you get the full source code and IP. Every launch comes with a 30-day post-launch warranty.

The biggest red flag in a quote for a meditation app is a single number with no hours behind it. Here are the others I'd look for.

• No hours per feature. If you can't see the player, downloads and subscription as separate lines, you can't tell what was left out.
• Offline downloads listed as a checkbox. Ask what happens to downloaded premium content when a subscription lapses.
• No admin tool. Ask how a non-engineer publishes a new session. If the answer is a ticket to the developer, budget for that forever.
• "Content library included." Ask how many sessions, who writes them, who records them and who owns the recordings.
• No mention of Apple's Small Business Program or Google's subscription fee in the revenue model.
• A web checkout with no annual reminder, no consent record and no web cancel path.
• Analytics SDKs on health events with no consent gate.
• Marketing copy that promises to treat insomnia or anxiety with no study behind it.
• Market-size or "X% of people are stressed" statistics with no primary source in the proposal.
• No mention of who owns the source code at the end.

A good quote can be wrong. A quote you can't check can't even be wrong, which is worse.

Several numbers that would sharpen this estimate are not public, and I've left them out rather than guess.

• Calm's engineering spend and headcount. It is private and files no 10-K.
• Calm's revenue, subscriber count and valuation. Figures online are third-party estimates.
• Narrator, celebrity and composer fees for meditation and sleep content. Not publicly disclosed.
• A published union rate for meditation or sleep narration. I found none I could verify.
• Calm's own infrastructure and CDN costs.
• Whether the FTC will propose a new negative option rule, and what it would say.

The Calm Help Center page on offline use blocked my automated fetch, so I relied on the search listing of that page for how downloads work. Every hour figure in this article is my scenario estimate. Your product will differ, and so will ours once we've seen your feature list.

Three steps, and none of them require hiring anyone.

• Write your feature list and cross out everything on the second list above. Then decide whether v1 needs offline downloads or can stream only.
• Write ten session scripts yourself and generate them with a synthetic voice for a few dollars. Play them to ten people in your audience before you pay a narrator.
• Draft your paywall copy with the trial length, the price after the trial and a cancel path, and read it against ROSCA and California's AB 2863.

Then take the list to any developer, us included, and ask for hours per feature. If they won't give you hours, you've learned something.

Time to get to work.