Transcription service cost: build your real annual bill

The price you should care about is your annual bill, not the per-minute rate on a vendor's homepage. A $0.006-per-minute API and a $20-per-seat subscription can produce wildly different yearly totals for the exact same workload, and the cheaper headline rate is very often the more expensive bill once you add up volume, seats, hidden multipliers, and your own time.

So stop comparing rates. Calculate totals. This guide gives you a worksheet to estimate your real annual transcription cost in five inputs, the verified 2026 rates to plug into it, a worked example, and the surcharges that quietly inflate the number. For recurring meetings, a flat per-seat AI tool almost always wins the math; for accuracy-critical deliverables, the human premium is worth it; and "free" native captions usually hide a license cost you already pay.

Why "cost per minute" is the wrong number to shop on

A per-minute rate only tells you the cost of one unit. Your bill is units times volume, plus everything the unit price excludes. Two things break the simple comparison.

First, the rates are attached to different units. Human services and raw APIs charge per audio minute. AI meeting tools charge per seat per month with minutes bundled (often unlimited). Native platform captions are "free" but bundled into a per-seat license you already buy. You cannot compare $1.99 per minute to $20 per seat without converting both into an annual total for your specific workload.

Second, the cheap rates exclude the most expensive parts. A raw speech-to-text API gives you text and nothing else: no capture off the call, no recording, no speaker labels, no summary, no storage, no search. Every one of those becomes a line item you pay for in other vendor spend or in engineering time. The 36-cent transcript is real; the workflow around it is not free.

The five inputs that decide your bill

Before you look at any rate, write down five numbers about how you actually transcribe. Everything else is arithmetic on top of these.

1. Hours per month. Add up the audio you transcribe in a typical month. For meetings, count the calls you attend, not just the ones you host. Most knowledge workers land between 5 and 40 hours.
2. Seats. How many people need their own transcripts and notes? One freelancer is one seat. A 12-person sales team is twelve.
3. Accuracy floor. Is the transcript an internal artifact (notes, action items, a searchable archive) or a deliverable where errors carry legal, medical, or published consequences? This single answer decides whether you ever touch human pricing.
4. Frequency shape. Steady recurring volume every week, or rare one-off files? Steady volume rewards flat subscriptions; one-off files reward pay-per-minute or a free tier.
5. What you need around the text. Just the words, or recording, speaker labels, summaries, search, and integrations too? The more of these you need, the worse a bare API looks and the better a bundled tool looks.

Hold those five numbers. The rest of this guide turns them into a dollar figure.

The verified 2026 rates to plug in

Here are the current, real numbers, each traced to a vendor pricing page as of June 2026. Use these as the rate inputs in your worksheet.

$1.99/min

Rev human, 99%+ accuracy

$0.25/min

Rev AI, 95%+ accuracy

$0.006/min

OpenAI gpt-4o-transcribe

$0.15/hr

AssemblyAI Universal-2

$29/seat/mo

Scribbl Pro, unlimited

$0/mo

Scribbl free, 10 meetings

Human services. Rev charges $1.99 per audio minute for professional human transcription at 99%-plus accuracy, with turnaround in 12 hours or less. GoTranscript advertises human work from roughly $1.02 to $1.58 per minute on its public estimator, with rush orders priced higher. That is roughly $60 to $120 for a one-hour recording before verbatim or rush surcharges.

AI meeting tools (per seat). Otter.ai Pro is $16.99 per user per month, or $8.33 billed annually, with 1,200 transcription minutes a month and a 90-minute meeting cap. Otter Business is $30 per user per month, or $19.99 annually, with unlimited meetings and a 4-hour cap. Rev sells subscription tiers too. Scribbl is free for up to 10 meetings a month at unlimited length, and Pro is $20 per user per month, or $13 billed annually, with unlimited meetings.

Raw STT APIs (per minute or hour). AssemblyAI async speech-to-text is $0.15 per hour on Universal-2. OpenAI lists gpt-4o-transcribe at an estimated $0.006 per minute and gpt-4o-mini-transcribe at $0.003 per minute. On paper these are the cheapest numbers here. They are also the most misleading, because they cover the text alone.

Native platform captions. "Free" only in a narrow sense: bundled into a paid Google Workspace, Microsoft 365, or Zoom license you already pay for, and gated behind manual activation and organizer-only ownership.

Run the worksheet: a copy-pasteable cost model

Here is the model. Fill in your five inputs from above, then compute the annual total for each option you are considering. Drop this into a spreadsheet or do it on paper.

ANNUAL TRANSCRIPTION COST WORKSHEET

Your inputs
  A. Hours per month .................... [HOURS]
  B. Seats (people needing notes) ....... [SEATS]
  C. Accuracy floor ..................... [INTERNAL] or [LEGAL-GRADE]
  D. Surcharge multiplier ............... 1.0 to 1.5  (see next section)
  E. Your hourly rate (for build/admin)   [$/HR]

Option 1: Human service (per minute)
  = A x 60 min x [RATE/MIN] x 12 months x D
  Example at $1.99/min, 10 hrs/mo:
  = 10 x 60 x 1.99 x 12 x 1.0  = $14,328 / yr

Option 2: AI meeting tool (per seat)
  = [SEAT PRICE/MO] x B x 12 months
  Example at $13/seat (annual), 1 seat:
  = 13 x 1 x 12             = $156 / yr
  (Often $0 on a free tier under the meeting cap.)

Option 3: Raw STT API (per minute) + build
  text   = A x 60 x [API RATE/MIN] x 12
  build  = (capture + storage + diarization + UI hours) x E
  Example text at $0.006/min, 10 hrs/mo:
  = 10 x 60 x 0.006 x 12   = $43 / yr   (text only)
  + build/maintenance time = the real cost

Option 4: Native captions (Meet/Teams/Zoom)
  = [SEAT LICENSE/MO already paid] x B x 12
  Marginal cost if you already own the license = $0
  But: manual start per meeting, organizer-only ownership

The arithmetic is deliberately blunt because the gaps are enormous. For one person transcribing 10 hours of meetings a month, the human-service option is over $14,000 a year, the per-seat AI tool is $156 (or $0 on a free tier), and the raw API text is $43 plus however many hours you spend building everything around it. Same workload, three orders of magnitude apart.

The surcharge multiplier nobody quotes you

Input D in the worksheet is the multiplier that turns a headline rate into a real bill. Standard pricing assumes the easy case. Here is what pushes it up:

• Turnaround. Standard human pricing assumes you can wait. Rush orders (6 to 12 hours) cost more, sometimes a lot more.
• Verbatim vs clean read. Capturing every "um," false start, and stutter takes longer and costs more than a cleaned transcript.
• Audio quality and speaker count. Bad audio and many overlapping speakers slow human transcribers and lower AI accuracy, which can push you up a tier.
• Multilingual and translation. Non-English audio and translated output usually carry surcharges. Microsoft Teams translated transcription, for example, requires the Teams Premium add-on.
• Caps and expiry. Otter's free tier caps at 300 minutes a month with a 30-minute-per-conversation limit; Rev Free is 45 AI minutes a month. Many tools also expire saved video or audio after a retention window, so "unlimited" can quietly mean "unlimited until it disappears."

For routine internal meeting notes on decent audio, D is about 1.0. For hard audio, rush turnaround, or verbatim legal work, D can reach 1.5 or higher. Price your real case, not the homepage case.

"Free" native transcription has a price tag too

This is the most expensive myth, because people build a workflow on "free" and discover the catch later. Native captions are bundled into licenses, not actually free.

So Google Meet transcription is "free" only if your organization already pays for Business Standard or higher, which is a per-seat cost in its own right. Even then it is opt-in per meeting and owned by whoever organized the call, which is a problem if you joined someone else's meeting or want your own copy. For the native walkthrough, see getting a Google Meet recording transcript. If you compare the full landscape of priced and free options, the four-bucket breakdown of what transcription costs sits alongside this worksheet.

Map your worksheet result to the right option

Run your five inputs through the worksheet, then route to the cheapest correct option. Two questions sort almost everyone.

1. 1

   Need 99% legal-grade accuracy?

   If yes, use a human service (Rev, GoTranscript) and budget it per deliverable. Your worksheet will be large; that is the cost of insurance. Stop here.

2. 2

   One-off or rare files?

   Use a free tier or pay per minute. Don't buy a subscription for volume you don't have.

3. 3

   Recurring meetings, just you?

   Use a flat per-seat AI meeting tool. A free tier (like Scribbl's 10 meetings a month) often means your worksheet total is $0.

4. 4

   A team of meeting-heavy users?

   Per-seat unlimited beats per-minute the moment you cross a few hours a month per person. Multiply seats, not minutes.

5. 5

   High-volume programmatic pipeline?

   Use a raw STT API, but add the build line to your worksheet before you commit.

The pattern repeats in every worksheet: heavy meeting users blow past per-minute math fast, so flat unlimited per-seat wins. Light or one-off users do better on a free tier or pay-per-minute. Accuracy-critical deliverables justify the human premium. The API route pays off only at programmatic scale with engineering already in place.

Recurring meetings

• Winner: flat per-seat AI tool
• Annual cost: $0 to ~$240/seat
• Why: bundled minutes, recording, summary, search
• Watch: meeting caps and video-retention expiry

Legal-grade deliverable

• Winner: human service (or AI draft plus human cleanup)
• Annual cost: scales with volume at ~$1 to $2/min
• Why: 99%+ accuracy is the deliverable
• Watch: rush and verbatim surcharges

Where Scribbl fits the worksheet

For the most common workload here, recurring Google Meet calls, the worksheet almost always lands on a flat per-seat AI tool, and that is the case Scribbl is built for.

Scribbl records, transcribes, and summarizes your Google Meet calls from the browser with no bot joining the meeting. The free tier covers 10 meetings a month at unlimited length, so an individual transcribing their regular calls frequently lands at $0 on the worksheet. Pro is a flat $13 to $20 per seat with unlimited minutes, which beats per-minute API math for anyone past a few hours a month and does not require a paid Business Standard Workspace edition to get a usable Meet transcript. For teams, the desktop app adds Zoom and Microsoft Teams along with CRM, Slack, and Drive integrations, so you avoid paying a separate per-platform license tax. Compare tiers on the pricing page.

To be honest about when Scribbl is not the answer: if you need a court-admissible verbatim transcript, your worksheet should point to a human service. If you run a high-volume programmatic pipeline, use a raw STT API. For everyone whose real job is turning meetings into notes and action items, a flat per-seat tool is almost always the cheapest line on the worksheet. If you want the no-cost angle specifically, the companion piece on the best free transcription software breaks down the $0 options by category.

Frequently asked questions

How much does transcription cost per hour in 2026?

It depends on the bucket. A human service runs roughly $60 to $120 an hour (Rev at $1.99 per minute, GoTranscript from about $1.02 to $1.58). A raw STT API is cents an hour (OpenAI gpt-4o-transcribe at $0.006 per minute is about 36 cents; AssemblyAI from $0.15), but excludes everything around the text. An AI meeting tool bundles that hour into a flat per-seat subscription ($0 to about $30 a month) regardless of how many hours you run.

How do I estimate my total annual transcription cost?

Use the five inputs in this guide: hours per month, seats, accuracy floor, surcharge multiplier, and your hourly rate for any build or admin time. Then compute (rate times annual volume) plus seats plus surcharges plus your time for each option. The fully loaded annual total is the only number that compares fairly across human services, AI tools, raw APIs, and native captions.

Is per-minute or per-seat pricing cheaper for meetings?

Estimate your monthly minutes per person. Occasional transcription is cheaper per minute or on a free tier. For recurring meetings, do the division: a $13-per-seat unlimited plan beats per-minute the moment you cross a few hours a month, and the gap widens fast for meeting-heavy users. Per-minute pricing punishes consistency.

Is AI transcription accurate enough to skip the human cost?

For internal use, yes. AI transcription typically hits 90 to 95% accuracy, which is fine for meeting notes, action items, and a searchable archive. It is not enough when the transcript itself is a legal, medical, or published deliverable, where 99%-plus human accuracy is worth the 8 to 10x premium. Match accuracy to stakes, and you stop paying for precision you do not need.

Is Google Meet, Teams, or Zoom transcription actually free?

Mostly no. Google Meet native transcription requires a Business Standard or higher Workspace edition and a manual start, and the transcript belongs to the organizer. Teams stores the live transcript in the organizer's OneDrive, and translated transcription needs Teams Premium. Zoom live captions are genuinely free for all accounts, but a saved transcript needs a paid plan plus cloud recording. The marginal cost is $0 only if you already pay for the right license.