1. Choose live text, retained audio, or file transcription
Direct dictation converts speech into editable text at the current note insertion point. Voice note and Meeting recorder begin with an audio recording and can show live transcription. AI Meeting Notes adds speaker labels and summaries to eligible recordings. AI Transcribe processes attachments or a one-off item through Evernote’s app or standalone web tool.
Those controls can all display recognized words, but they do not save the same artifact. Decide whether the primary result should be text in the note, a retained recording, a meeting transcript with speakers, or a transcription created from existing media before selecting the microphone or upload control.
- Use direct dictation for editable prose at the cursor.
- Use Voice note when the recording should stay with the note.
- Use Meeting recorder or AI Transcribe for longer captured or existing media.
2. Use Evernote’s direct dictation for live note text
Evernote’s current 11.21.5 release notes say desktop and web users can dictate directly into a note and convert speech to text. Open the exact notebook and note, focus the title or body position, begin the native dictation control, speak, stop, and inspect the text before adding formatting, links, tasks, mentions, or shares.
The release note does not currently name the direct-dictation shortcut, speech provider, processing location, audio retention, supported languages, or plan boundary. Treat those details as undisclosed for this specific control instead of copying the documented behavior of AI Transcribe, AI Meeting Notes, Apple Dictation, or another speech tool.
3. Keep Voice note and Meeting recorder as audio workflows
Evernote 11.21.5 redesigns audio recording into Voice note and Meeting recorder modes and can display live transcription in a floating, draggable panel. The recording remains the captured artifact inside the Evernote note even when recognized text is visible during the session.
Use Voice note for a personal spoken capture and Meeting recorder when the conversation, speaker separation, or later summary matters. Confirm the note, notebook, microphone, participants, local consent requirements, and whether audio should be retained before recording.
4. Use AI Meeting Notes for speakers, transcript, and summary
Evernote AI Meeting Notes can begin from a new recording or an uploaded audio file. On desktop and web, select Transcribe on the recording; before processing, use the dropdown to enable or disable speaker recognition. Evernote can label speakers, rename each speaker across the transcript, regenerate with another language or speaker setting, and create a summary after enough text exists.
The current guide lists a 200 MB or 120-minute limit for the newer audio-only Meeting Notes experience, requires at least 250 transcript characters for a summary, and says AI processing needs an internet connection even though recording can begin offline. Keep the note open while processing. The transcript and summary cannot currently be downloaded directly, but the transcript can be copied; the recording can be downloaded.
Evernote’s current guide still labels AI Meeting Notes a beta preview for paid users while saying it will roll out more broadly as part of V11. Check the active account and current product state rather than promising universal access.
5. Use AI Transcribe for attachments or a one-off web job
Inside Evernote, attach an audio, video, or image file to a note. Current audio files are transcribed automatically below the attachment; images and videos expose a Transcribe control. The standalone AI Transcribe web tool also accepts supported files and URLs, then offers Copy, text download, or Save to Evernote.
The current AI Transcribe FAQ lists audio and video formats including MP3, MP4, M4A, WAV, WebM, AAC, MOV, OGG, MKV, and M4V, plus common image formats and supported web URLs. It lists a 100 MB or 60-minute media limit and more than 50 languages. Those limits belong to the general AI Transcribe surface, not the newer 200 MB or 120-minute audio-only Meeting Notes path.
6. Keep each Evernote processing statement at its source
Evernote says audio transcription processing occurs within Evernote. It separately says audio summarization plus image and video transcription can involve a third-party AI vendor, the content is used only for the requested processing, it is not used for model training, and third-party processors delete files within 30 days.
Evernote’s broader AI FAQ says AI Meeting Notes uses Evernote internal models for audio transcription while audio summaries may use a third-party vendor. Do not transfer either statement to direct in-note dictation, system Dictation, a standalone URL, saved recordings, copied transcripts, synced notes, or every AI feature unless the current documentation explicitly covers that control.
7. Review note sync, sharing, and history separately
Evernote automatically saves and syncs notes across connected devices. Text produced by any input path becomes Evernote account content once it is in the note, and retained recordings remain attachments. Check the notebook, note title, tags, links, tasks, people mentions, and sharing state before assuming the content is private.
If speech overwrites or removes important text, use Undo immediately. Evernote Note history can preview and restore periodic older versions on Mac, Windows, or web, but it is not a per-word safety guarantee and is separate from the audio, transcript, and processing controls.
8. Troubleshoot the path that actually failed
If direct dictation is missing, confirm Evernote 11.21.5 or later, desktop or web availability, the current account rollout, microphone permission, and a focused editable note. Because the release note does not publish one shortcut or provider contract, do not troubleshoot it as if it were Apple Dictation.
If Transcribe is missing, confirm AI Transcription is enabled under Settings → Preferences → AI features, the attachment is synced, the format and size fit the correct path, and the account has current access. If processing stops, keep the note open, restore an internet connection, retry, refresh, or reopen the app. Test macOS Dictation separately in TextEdit when only the system input path fails.
9. Add IraVoice only for a distinct local text path
Use Evernote’s direct control first for native in-note speech, Voice note or Meeting recorder when audio should stay in Evernote, and AI Transcribe for supported recorded media or URLs.
IraVoice 0.7.1 is a separate option for macOS 26+ Apple-silicon Macs when recognition and supported formatting should run on the Mac across supported focused text fields, or when speech should become reviewable Markdown before entering a note. IraVoice does not sign in to Evernote, choose notebooks, start or retain recordings, identify speakers, summarize meetings, upload files, transcribe attachments, use Evernote AI, change sharing, restore history, or create and sync notes.