You have a file from Teams, Zoom, Meet or a dictaphone and need text out of it. Nothing to install, and nobody joins the call.
You upload the meeting recording afterwards: a file from Teams, Zoom, Meet or a dictaphone, with nothing to install on the call itself. You get the whole thing with timestamps and speaker labels, search across the text, and TXT and SRT export. A 90-minute meeting costs $9.00 as a one-off. Recordings up to five minutes are free once you sign in.
Drop a file here or choose one from your device. We will show you the beginning of the transcript first.
skryba.ai transcribes on Scribe v2, which gets 2.2% of words wrong in Artificial Analysis' independent index — the lowest of every model measured. Gemini 3.5 Transcribe gets 2.6% wrong, Whisper large-v3 4.1%. It is the same model that writes the transcript preview you see before paying.
| Model | WER |
|---|---|
| skryba.ai (Scribe v2)the model we transcribe on | 2.2% |
| MAI-Transcribe-1.5Microsoft Azure | 2.4% |
| Gemini 3.5 TranscribeGoogle | 2.6% |
| Universal-3 ProAssemblyAI | 3.1% |
| GPT TranscribeOpenAI | 3.3% |
| Whisper large-v3OpenAI | 4.1% |
| Nova-3Deepgram | 5.2% |
Source: Artificial Analysis, AA-WER v2 index, non-streaming mode — read 27 August 2026. Seven of the 40-plus models measured: the top of the table and the names you already know. The index is measured mostly on English audio; for other languages, Polish included, no comparable public ranking exists. That is why we show you the start of every transcript before we ask for anything.
Most meeting tools ask you to let an assistant into the call. We work the other way round: the call ends, you upload the file, we hand back the text. Nobody joins the meeting, nobody shows up on the participant list, and there is nothing to explain to a client.
There is one condition, and we say it up front: you need a recording. If the meeting was not recorded there is nothing to make text from — recording is started by the platform or by a dictaphone on the table, not by us.
If what you want is an assistant that joins the call by itself and boils it down to bullet points, this is not that tool, and creating an account here would waste your time.
The second thing worth knowing before uploading anything: Teams, Zoom and Google Meet have their own transcription, switched on during the meeting, included in most business plans. If the meeting is still ahead of you and runs on one of those platforms, switch it on there — there is nothing to upload. skryba.ai is for when the recording already exists as a file: a meeting recorded by a dictaphone on the table, a conversation someone forwarded to you, or a platform where nobody turned transcription on.
Every platform stores the recording on its own side and lets you download it. You upload the file exactly as you received it — no conversion needed, and we pull the audio track out of video on our side.
| Source | What you usually get | What to do |
|---|---|---|
| Microsoft Teams | MP4 from the cloud, sometimes audio alone as M4A | Upload it unchanged |
| Zoom | An MP4 plus a separate M4A audio file | Pick the M4A — smaller, same result |
| Google Meet | An MP4 saved to Google Drive | Download from Drive and upload |
| A dictaphone or a phone on the table | MP3, WAV, WMA or M4A | Copy it off and upload |
| A browser-tab recording | WebM | Upload it unchanged |
Scroll the table sideways to see the remaining columns.
We will not fetch a meeting link or a YouTube address. A direct HTTPS address to the file itself does work — from a company drive that exposes one, say — and those we fetch up to 500 MB.
Minutes are written from that record rather than instead of it: you have the exact wording of what was agreed and the time it was said, so you never scrub through the audio hunting for one sentence.
You pay per started minute: $0.10 a minute, with a $1.99 minimum. An hour-long meeting is $6.00, a 90-minute one is $9.00. If there are several such meetings a month, buying hours up front works out cheaper.
| How often you record meetings | What to choose | Cost |
|---|---|---|
| Now and then, one recording | A single transcript | $0.10 a minute, from $1.99 |
| A few meetings, but irregularly | The 5-hour pack, no expiry | $19.99 ($4.00 an hour) |
| A weekly status call and a few client calls | The Starter subscription, 4 hours a month | $9.99 a month ($2.50 an hour) |
| A dozen-plus hours of recordings a month | The Pro subscription, 10 hours a month | $19.99 a month ($2.00 an hour) |
Scroll the table sideways to see the remaining columns.
Subscription hours expire at the end of the billing period; pack hours stay on the account until you use them. There is also a 10-hour pack at $29.99 if you would rather buy once and forget about it. For 14 days after buying you can withdraw and get the whole amount back, including for hours already used.
A recording of a conversation is personal data about everyone who spoke on it. The controller is you or your company; we process the file in order to return text. Below are the facts you can build on.
The cleaned-up version removes fillers and stammers from the record, and only those. It does not remove names or any other personal data — if the minutes have to go out anonymised, you change those passages in the editor before exporting.
Yes, with no conversion. Teams and Zoom hand back an MP4, and Zoom also an audio-only M4A; Meet saves an MP4 to Google Drive, and a browser-tab recording is usually WebM. You upload all of those whole — we pull the audio track out of video on our side, so there is nothing to export first.
Yes, speaker recognition is on by default and you name the speakers in the editor. On a meeting where people speak in turn the split is good. With six people talking over each other, or one microphone in the middle of the room, it will need corrections. The free preview shows the beginning of the record together with the split, so you can check it on your own file.
You pick the recording language before transcription, or leave automatic detection on. For a conversation held in Polish with English names and acronyms dropped in, pick Polish. For a meeting that switches between two languages, leave detection on automatic and judge the result on the free preview — this is the case where seeing the text before paying is worth the minute it takes.
Usually a few minutes. You do not have to sit through it: close the page and the transcript waits on your account. With a large video file the upload itself takes longest, which is why an M4A version of a Zoom recording is worth choosing when the platform offers both.
Into a private bucket on Cloudflare R2 within the European Economic Area. We keep it until you delete it, or until the auto-delete setting removes it after 1, 7 or 30 days — the transcript survives that deletion. You can delete a recording together with its transcript and shares by hand at any time, and the same goes for the whole account.
Not for the preview — you upload the file and read the beginning of the transcript without giving an email address. You can also buy a single meeting with no account: pay by card or BLIK, read the whole thing, and the recording stays available for 30 days. An account is needed for free recordings up to five minutes, and for keeping every record in one place.
