Communication · United States · USD

Cost to build an app like Zoom

A Zoom-like app is a media infrastructure product. Two people on a call is a solved problem; fifty people on a call with screen sharing, recording and someone on hotel wifi is not. Selective forwarding, simulcast, bandwidth adaptation and graceful degradation are where the budget goes.

What you are actually building

An app like Zoom is three or four screens people recognise sitting on a great deal of engineering they never see. These are the screens that define the shape — and the part of each one that costs real money.

  • The call grid

    WebRTC and an SFU, layouts from two to fifty participants, active-speaker switching. A peer mesh works for three people and dies at five.

  • Quality under load

    Bandwidth estimation, simulcast and packet-loss concealment. Call quality is the product; everything around it is chrome.

  • Share and record

    Screen share, cloud recording and transcription — server-side media processing, billed by the minute, forever.

Price your own version

The calculator below is preloaded with the Zoom template — its features, architecture and a planning scale for year one. Remove what you do not need, add what you do, and the estimate follows.

Building an app like Zoom

A starting template, not a specification. Every feature, number and technology choice below is yours to change.

Reference template

Zoom

A Zoom-like app is a media infrastructure product. Two people on a call is a solved problem; fifty people on a call with screen sharing, recording and someone on hotel wifi is not. Selective forwarding, simulcast, bandwidth adaptation and graceful degradation are where the budget goes.

Preloaded features
12
Category
Productivity
  • The call grid

    WebRTC and an SFU, layouts from two to fifty participants, active-speaker switching. A peer mesh works for three people and dies at five.

  • Quality under load

    Bandwidth estimation, simulcast and packet-loss concealment. Call quality is the product; everything around it is chrome.

  • Share and record

    Screen share, cloud recording and transcription — server-side media processing, billed by the minute, forever.

What makes a Zoom-like app expensive

Three things dominate this shape of product. Everything else is comparatively ordinary app development.

The media server

An SFU handling many participants with simulcast and adaptive layer selection. Self-hosting is cheaper at volume and considerably more work than using a vendor.

Quality under bad networks

Packet loss concealment, dynamic bitrate, audio priority and reconnection. This is the difference between a demo and a product.

Recording and transcription

Server-side composition, storage, playback and increasingly transcripts and summaries.

Cost breakdown for the default template

Before you change anything. Costed at the midpoint of a $118–$155/hr blended US rate.

$846K6,197 hours
  • Product planning$27,000199 hours · 3% of the buildDiscovery, requirements, user stories, technical scoping.
  • UI/UX design$40,000292 hours · 5% of the buildScreens, design system, prototypes, usability and accessibility work.
  • Mobile development$89,000651 hours · 10% of the buildThe app itself — iOS, Android or a shared cross-platform codebase.
  • Backend development$200,0001,463 hours · 24% of the buildAPIs, business logic, data model, admin and background processing.
  • Infrastructure engineering$116,000850 hours · 14% of the buildScaling, caching, queues, storage, media pipelines, observability.
  • Third-party integrations$48,000355 hours · 6% of the buildBuilding and hardening each external service connection.
  • Security & compliance$113,000830 hours · 13% of the buildControls, hardening, audit logging, review and testing.
  • QA & testing$98,000716 hours · 12% of the buildFunctional, regression, device-matrix, performance and release testing.
  • DevOps & release$35,000254 hours · 4% of the buildCI/CD, environments, infrastructure as code, monitoring, store releases.
  • Project management$80,000589 hours · 9% of the buildPlanning, coordination, reporting and the cost of keeping it on track.

Features in a Zoom-like app

The starting list, longest build first — each bar is that feature’s share of the biggest one, so you can see where the hours actually go. Every row is a checkbox in the calculator above: this is a template, not a specification.

Core features (10)

504h total
  • Admin dashboard129h
  • Subscriptions & plans89h
  • Group chat62h
  • External calendar sync61h
  • Roles & permissions47h
  • User profiles32h
  • Push notifications30h
  • Transactional email20h
  • Social login19h
  • Email & password login15h

Advanced features (2)

377h total
  • Large group meetings235h
  • Video calling142h

Integrations (6)

  • Stripe
  • Agora / Daily / LiveKit (video)
  • SendGrid / Resend (email)
  • AWS
  • Amplitude / Mixpanel
  • Sentry / Datadog (monitoring)

How long, and who builds it

The $658,000$1,076,000 figure assumes this schedule and this team. Both scale with the answers in the calculator above — and neither is a quote.

Delivery timeline — 9–12 months

  • Discovery & product definition4–5 wks
  • UI/UX design6–8 wks
  • Development (app + backend)20–28 wks
  • Integrations & infrastructure5–7 wks
  • QA & security testing5–7 wks
  • Store submission & launch2–3 wks

The team — 7 people

  • Product manager1×
  • UI/UX designer1×
  • Mobile developer1×
  • Backend developer1×
  • DevOps / SRE1×½
  • QA engineer1×
  • Security engineer1×

The phases add to more than the total on purpose: design starts before discovery finishes and QA runs alongside development rather than after it. 7 people is about 6.5 full-time equivalents — a smaller team takes proportionally longer, and a much larger one does not go proportionally faster. The two roles most often missing from a first plan are DevOps and QA, which together are 16% of this build.

Technology considerations

A stack that would serve this product shape well. A recommendation for what you could build — not a claim about what Zoom runs.

Suggested technology stack for a Zoom-like app
LayerSuggestion
MobileNative iOS and Android for media performance, plus a browser client via WebRTC
BackendGo or Node.js signalling services alongside an SFU media layer
DatabasePostgreSQL for accounts, meetings and entitlements
CacheRedis for room and participant state
Real-timeWebSocket signalling plus WebRTC media
CloudAWS
Object storageS3 for recordings and transcripts
CDNCloudFront for recording playback
Media pipelineSFU (LiveKit, Janus or a vendor) with simulcast, recording and transcription
MonitoringDatadog with per-call quality (jitter, packet loss, MOS) dashboards

Top cost drivers

  1. 1

    Medium scale traffic and usage18% of the build

    Caching, queues and real monitoring become mandatory. That is real engineering: caching layer (redis or equivalent) in front of hot reads; background job queue for anything slower than a request; read replicas and query-level database tuning. Lower your year-one traffic assumptions and this shrinks immediately.

  2. 2

    Advanced backend architecture10% of the build

    Needed when traffic is high enough, or spiky enough, that a single service cannot absorb it and a slow endpoint takes the whole product down with it.

  3. 3

    6 third-party integrations6% of the build

    Every external service is code you write, failure modes you handle and an API that will change without asking you. The ones you can defer are the cheapest scope cuts available.

  4. 4

    SOC 2 Type II compliance work5% of the build

    Engineering and evidence work to make an audit survivable. Certification fees, legal review and auditor time are separate costs that this estimate excludes.

  5. 5

    Cross-platform + web app3% of the build

    Two separate native codebases means the client is built twice. Moving to React Native or Flutter would cut roughly a third of your mobile hours, at the cost of native modules wherever you touch device hardware.

Infrastructure and maintenance

Neither of these is part of the build figure, and neither of them stops.

Cloud infrastructure

$1K $8K/month

  • Baseline environment$120–$420
  • Application compute$270–$960
  • Database$7–$27
  • Object storage$105–$130
  • CDN & bandwidth$282–$635
  • Media transcoding$250–$667

Annual maintenance

$99K $269K/year

15–25% of build cost per year: bug fixes, iOS and Android version updates, SDK and dependency upgrades, security patching and small improvements. Cloud costs are the separate figure alongside.

Building an app like Zoom — common questions

How much does it cost to build an app like Zoom?

Using this site's default profile — 12 features, cross-platform + web app, a advanced backend and medium scale traffic — the estimate is $658,000 to $1,076,000 for a US team, or roughly 6,197 engineering hours at a blended $118–$155 per hour. That figure covers design, both apps, backend, infrastructure, integrations, security, QA, DevOps and project management. It excludes cloud hosting, vendor fees, marketing and legal. Change any answer in the calculator on this page and the number moves with it.

How long does it take to build an app like Zoom?

About 9–12 months with the team this estimate assumes — 7 people, roughly 6.5 full-time equivalents. Discovery and design run first and partly in parallel, development takes about half the calendar, and QA overlaps the back half rather than following it. A smaller team takes proportionally longer; a much larger one does not go proportionally faster, which is why compressing this schedule costs more than it saves below a certain point.

What team do you need to build an app like Zoom?

1× product manager, 1× ui/ux designer, 1× mobile developer, 1× backend developer, 1× devops / sre (part-time), 1× qa engineer, 1× security engineer. That is what 6,197 hours of work distributed across 10 disciplines actually requires. The roles people most often leave out of their own plan are DevOps and QA, which together are about 16% of this build.

What does an app like Zoom cost to run each month?

Cloud infrastructure is estimated at $1,394 to $8,290 per month at the scale this profile assumes, driven mainly by media delivery and transcoding. Third-party services add $100 to $5,000. Separately, budget $99,000 to $269,000 a year for maintenance — bug fixes, OS and SDK updates and security patching. None of these are part of the build figure.

Should I build the media layer or buy it?

Buy it first. LiveKit, Daily or Agora handle the SFU for a usage fee, saving 400+ hours. Self-hosting starts to pay back at sustained high minute volumes, and by then you will know your actual usage patterns.

How many participants can one call support?

With a vendor SFU, 50–100 with video is routine and hundreds are possible with active-speaker limits. Beyond that you move to a broadcast model, which is a different architecture and a different cost profile.

Related app cost estimates