Audio
Configuration reference: Áudio
Overview
The Audio component stores a recording on the record: recorded on the device itself (computer, phone or browser) or imported from a file. It suits visit notes, dictation, fault reports or anything quicker to say than to type.
Optionally, every audio is transcribed with AI as soon as it is saved, and the text shows in the field itself, below the player, ready to read, select and copy.
Audio is stored in the same attachments storage as File and counts towards the plan's storage limit.
Recording and importing
- Gravar (Record) asks for microphone access the first time. While recording you see the elapsed time and the limit; Parar (Stop) finishes and uploads, Cancelar (Cancel) discards.
- Importar áudio (Import audio) accepts
webm,m4a,mp3,wavandoggfiles, up to 25 MB. - Recording stops on its own at the configured maximum duration (30 minutes at most).
- Without an internet connection you can still record, but the upload fails: the recording stays in the field with Tentar de novo (Try again) until there is a network, or it can be discarded.
- Closing the record with a recording still waiting to upload asks for confirmation before discarding it.
With audio saved, the field shows the player (play, pause and seek on the bar, with the audio's duration) and the button to remove the audio.
Replacing or removing the audio deletes the old file from storage when the record is saved.
AI transcription
With Transcrever (Transcribe) on, the transcription request goes out as soon as the audio finishes uploading, even on a new record that has not been saved yet:
- while it runs, the field shows A transcrever… (Transcribing…); the rest of the form keeps working;
- the text stays in the field itself, below the player: it can be selected and, when long, opens collapsed with Mostrar tudo (Show all);
- Copiar texto (Copy text) copies the transcript; with a Coluna do texto (Text column) chosen, Copiar para (Copy to) appends the text to that column of the same record, on a new line, without deleting what was there (it stays unsaved, like any edit);
- Transcrever de novo (Transcribe again) is always available: if the text did not come out right, it can be requested again, until it is right. If the transcription fails (no network, no credits, no speech in the audio), the field says why and shows Tentar de novo (Try again);
- with Preencher a coluna automaticamente (Fill the column automatically) on, the text is appended to the Coluna do texto by itself when the transcript arrives;
- Língua (Language) can stay on Detetar (Detect) or be fixed, which helps with short audio;
- each transcription spends the organization's AI credits, per second of audio, and shows in the portal's AI usage. Every new request spends again. Without credits the audio is still saved.
Offline, recording keeps working: the transcription is marked as failed and can be requested again once there is a network.
Audio that arrives without a transcript by other routes (automations, API) is transcribed on the server, a few seconds or minutes later: the text still lands in the field and only goes to the Coluna do texto with Preencher a coluna automaticamente. The write and the AI usage are made in the name of the person who saved the record or, with no known author, the organization's system user.
Replacing the audio starts a new transcription; removing the audio takes its transcript with it, but does not delete what was already copied to another column.
Configuration
| Setting | Description |
|---|---|
| Gravar no dispositivo | Shows the Record button. On by default. |
| Importar ficheiros | Shows the Import audio button. On by default. |
| Duração máxima | From 1 to 30 minutes; recording stops on its own here. |
| Texto de dica | Optional sentence shown in the empty field. |
| Transcrever | Turns on AI transcription: each audio is transcribed as soon as it is saved and the text stays in the field. |
| Coluna do texto | Optional: Text or Description column of this table to copy the text into. |
| Preencher a coluna automaticamente | With a Text column, appends the text to it by itself when the transcript arrives. Off on a new column; columns set up before this option keep copying. |
| Língua | Detect (default), Portuguese, English, Spanish, French, German, Italian or Dutch. |
Display in views
| View | Display |
|---|---|
| Form | Record and Import audio buttons; with audio, the full player, the remove button and, with transcription on, the text and the buttons to copy it or request it again |
| Table, List, Calendar, Kanban | Play button and the time, no bar |
Filters
Audio is not compared: filters only tell empty from not empty.
Stored value
The value is the same JSON as File for a single file, plus the duration in milliseconds and, with transcription on, the transcript (status is pending, done or failed; when it fails, error says why):
{
"path": "feature_<table>/column_<column>/1760798400000000_Gravacao_2026-10-18_14.05.m4a",
"fileName": "Gravação 2026-10-18 14.05.m4a",
"mimeType": "audio/mp4",
"fileSize": 482133,
"durationMs": 61234,
"syncStatus": "synced",
"transcript": {
"status": "done",
"text": "Customer visit: the pump filter needs replacing.",
"language": "en",
"at": "2026-10-18T14:06:02.000Z",
"requestId": "8c1f0d3e9b5a4f7e"
}
}
Permissions
The same as File: whoever edits the record records, imports, removes the audio and requests the transcript; whoever only reads listens and copies the text.