Sprecho can transcribe a recording you already have — a dictaphone file from a site visit, a voice memo from your phone, a recorded meeting — instead of only what you say live. Open Import audio in the sidebar, drop the file in, and the finished text goes into your history.
Audio import is part of the Command module. The Import audio entry appears in the sidebar only with a Command seat. Command is released in mid-September 2026; until then the import is not yet available.
What you can upload
- Formats: WAV, MP3, M4A and OGG.
- Size: up to 200 MB per file.
- Length: up to four hours per file.
- There is a minimum too: a recording of a fraction of a second is rejected as too short to transcribe.
The format is decided by reading the file, not by trusting its name. An
.mp3 that is really something else is refused with a clear message rather than
failing halfway through, and a correctly formatted file with the wrong extension
still works.
How to import a recording
- Open Import audio in the Sprecho sidebar.
- Drag the file onto the drop zone, or click Choose a file.
- Set Language — leave it on Detect automatically unless you already know what is on the recording.
- Set Style if you want the transcript formatted more or less formally.
- The upload starts immediately. You can keep using Sprecho while it runs.
The card then walks through its stages: Uploading → Waiting → Reading the file → Transcribing → Saving → Done. Long recordings are transcribed in parts, and the card counts them ("Transcribing part 3 of 12") so you can see progress on a two-hour file.
Cancel stops a running import. Remove clears a finished card from the list.
Where the transcript lands
When the import finishes, the card says Saved to your history. From there:
- Copy text puts the whole transcript on your clipboard.
- The transcript is searchable in your history alongside everything you have dictated live — see Notes and dictation history.
If your privacy settings switch off transcript storage, the import card says so and the text is not saved anywhere. Copy it before you navigate away — it will not be there later. See Turn off transcript and audio storage.
The list of import cards itself is temporary and is cleared when the server restarts. The transcripts stay in your history.
Limits worth knowing before you start
- Two imports at a time. Start a third and Sprecho asks you to wait for one to finish or cancel it.
- Words count against your quota exactly like live dictation, which is why a three-hour meeting is a paid capability rather than a free one.
- Availability is checked before the upload, not after. If your plan does not include audio import, the page says so at the top and no bytes are sent.
Getting a good transcript from a recording
Recordings you did not make for transcription are harder than live dictation, and a few things help:
- A single speaker close to the recorder transcribes far better than a meeting recorded from the middle of a table.
- Trim the silence at the start and end if your recorder captured minutes of nothing.
- Split very long meetings into logical chunks. A one-hour section is easier to review than a four-hour wall of text.
- Add names to the dictionary first. Participant and company names are the most common errors in meeting transcripts, and the dictionary applies to imports exactly as it does to live dictation.
Error messages and what they mean
| Message | What to do |
|---|---|
| That file is not a WAV, MP3, M4A or OGG recording | Convert it, or check that the file downloaded completely |
| That file contains no audio track | It is probably a video container with the audio stripped, or an empty export |
| That file could not be decoded | It is damaged or only partly downloaded — re-export it |
| That recording is longer than the import limit | Split it into pieces under four hours |
| You already have an import running | Wait for the current one, or cancel it |
| You have used up this week's words | The quota resets Monday |