Voice Capture
Record your voice and turn it into polished documents — dictate from a dedicated page or transcribe directly inside the editor.
Voice Capture lets you speak instead of type. Record your thoughts, and Opisense transcribes them in real time and creates a formatted document. It works for solo voice notes, meetings with multiple speakers, interviews, and dictation. There are two ways to use it: the standalone Voice Capture page, or the microphone button inside any document.
Two ways to capture
Voice Capture page
Open Voice Capture from the sidebar, choose an environment preset, and hit record. You see a live transcript as you speak — words appear in real time so you can follow along. When multiple people are talking, each speaker is automatically identified and color-coded. When you stop, the recording is processed and a polished document appears in Your Content.
In-document transcription
Click the microphone icon in the editor toolbar to start a live transcript block right where your cursor is. The transcript stays in place as part of your document — no separate page or extra steps.
Voice Capture is also available on the mobile app with the same real-time transcription — plus background recording and crash recovery for capturing ideas on the go.
What happens to your recording
When you use the Voice Capture page, your raw transcript doesn't stay raw. After you stop recording, Opisense automatically cleans it up:
- Filler words removed — "um," "uh," "like," and other verbal fillers are stripped out
- Grammar and punctuation corrected — Sentence structure is tidied up so the text reads naturally
- Organized into paragraphs — The wall of text is broken into readable sections based on topic flow
- Speaker-attributed meeting notes — When multiple speakers are detected, the AI produces structured meeting notes with speaker labels, topic grouping, and action items
- Original audio saved — The audio recording is stored alongside the document, so you can always go back and listen
AI cleanup applies to the standalone Voice Capture page only. In-document transcription keeps the raw transcript in place so you can edit it yourself.
Multi-speaker support
Voice Capture automatically detects when more than one person is speaking. During recording, each speaker is assigned a color and label (Speaker 1, Speaker 2, etc.) so you can follow who's saying what in real time. When the recording is processed, the AI uses this information to create structured meeting notes with speaker attribution — perfect for meetings, interviews, and group discussions.
If only one person is speaking, Voice Capture behaves as a single-speaker recorder with no speaker labels — exactly the way it works for solo voice notes and dictation.
Environment presets
Different recording environments need different sensitivity settings. Voice Capture offers three presets that adjust how aggressively background noise is filtered:
| Preset | Best for |
|---|---|
| Quiet | Home office, solo dictation — more sensitive, catches soft speech |
| Office | Normal office environment — balanced filtering |
| Noisy | Cafe, open floor plan — aggressive filtering, needs clearer speech |
Select a preset from the pill buttons below the record button. The default is Office. Your choice is saved and persists between sessions.
Supported languages
Voice Capture supports 10 languages for speech recognition:
| Language | Code |
|---|---|
| English | eng |
| Norwegian | nor |
| Swedish | swe |
| Danish | dan |
| German | deu |
| French | fra |
| Spanish | spa |
| Dutch | nld |
| Italian | ita |
| Portuguese | por |
Select your language from the small dropdown in the top-right corner of the recording area. The default is English.
Next steps
Recording
Use the Voice Capture page to record, transcribe, and create documents from your voice.
In-document transcription
Start a live transcription directly inside the editor without leaving your document.
Troubleshooting
Fix common issues with microphone permissions, browser support, and recording errors.