$3, Once, And You Keep It If We Fail: Why We Never Built a Transcription Subscription
We didn't design our pricing in a meeting. We designed it by reading complaints, and almost all of them were about subscriptions on *other* tools.
The same four stories kept arriving. Someone needed one interview transcribed, signed up for a $20 to $30 a month plan, and used it exactly once. Someone else forgot to cancel and got billed for four months of a tool they touched in January. Another hit a “minutes per month” cap nobody had flagged at checkout, mid-project, on a deadline. And almost everyone, at some point, asked the same plain question: *can I just pay for the thing I need?*
So that's what we built. **One transcription costs $3. One time. One credit, one job.** Your first one is free. There's no plan to forget to cancel, because there is no plan.
What a “balanced” article would tell you here
The standard take is that subscriptions reward heavy users and pay-as-you-go rewards light users, so “it depends.” That framing is wrong about what actually annoys people. The people who wrote to us weren't doing cost-per-minute math. They were angry about being charged for *nothing*: for idle months, for forgotten renewals, for caps invented to push them up a tier. A subscription is a bet that you'll lose track. The complaints were really about that bet.
The part nobody else does: we don't take the credit when *we* fail
Here's the rule that matters most, and it's written into our pricing page: **a credit is only consumed when a transcription completes.** If a job breaks during processing, whether that's a bad file, a model error, or anything else on our side, the credit stays on your account until you get a finished transcript out of it.
Think about what the subscription version of that failure looks like: the clock on your month keeps running whether the tool worked or not. You paid for time, and time passes regardless. A credit that survives failure flips the incentive. We only get to keep your $3 if you got your transcript. “What you see is what you get” isn't a slogan here; it's the literal accounting.
And now: an editor that never sees your file
When we added a transcript editor, the obvious move was a cloud editor: upload, edit, re-download, and let us hold the file. We refused.
Our new editor at /editor is free and runs entirely in your browser. You paste or open a transcript, fix the speaker names, merge segments, and correct the text, then export to TXT, Markdown, CSV, SRT, VTT, or Word. None of it touches a server. We even wrote our own DOCX exporter from scratch, so generating a Word file doesn't require shipping your text off to some library's API. A `.docx` is really just a zipped bundle of XML, and your browser can assemble it locally.
For anyone editing a transcript of a medical call, a legal deposition, a confidential interview, or an HR conversation, this is the whole game: the safest place for sensitive audio is a tool that never asks for it. Nothing should ever have to leave your device just to fix a few names.
The lesson
Pricing and privacy are the same signal wearing two outfits: both come down to what a company bets on. A subscription bets you'll forget to cancel. A cloud editor bets you won't mind handing over the file. We took the other side of both bets: a $3 credit that survives our own failures, and an editor that can't leak your data because it never receives it.
When this doesn't apply: if you genuinely transcribe high volume every single day, a flat monthly rate might work out cheaper, and we'll say so. We're not the cheapest tool for a transcription factory. We're the honest one for everyone else.
Try it
Open the free editor: no account, no upload, nothing to cancel. When you have audio to transcribe, your first credit is on us.