1. Choose the artifact before starting a microphone
Use Mac Dictation when you want editable words in an article body, title, subtitle, Note, comment, or other supported text field. The output is ordinary Substack text at the focused insertion point.
Use Voiceover when readers should hear the author read a text post, an audio embed when sound belongs at a chosen position inside a post, a podcast post when the recording is the primary artifact, or text-to-speech when a reader wants the Substack app to read an eligible English post. These paths differ in storage, placement, playback, permissions, and publication behavior.
2. Dictate into a Substack article on Mac
Sign in to the intended publication, open its Dashboard, choose Create → Article, and click in the exact body position where the words belong. Trigger the shortcut currently shown under System Settings → Keyboard → Dictation, speak a short passage, stop, and inspect the visible draft.
Apple says Dictation can enter text anywhere you can type, but the destination still owns its editor behavior. Add or verify headings, links, embeds, paywalls, code, names, quotations, and formatting after the prose is correct. Do not assume a spoken formatting phrase created the intended Substack structure.
- Confirm the publication and draft before a long session.
- Keep the insertion point in the article body.
- Review names, links, quotations, punctuation, and paragraph breaks.
- Stop Dictation before discussing the draft or changing fields.
3. Dictate a Note without confusing it with a newsletter
Substack Notes are short-form public posts that can contain text, links, images, quotes, and comments. Open Create → Note on the web, focus the composer, dictate the text, stop, and review it before choosing Post or Schedule.
Substack says Notes can appear on the web, in the app, and on a profile. Subscribers are not emailed for each new Note as they are for newsletters, although Notes may appear in a roundup or push notification. A working text cursor does not choose visibility, add media, schedule the Note, or publish it.
4. Treat Continue and Publish as separate decisions
Substack automatically saves post drafts and exposes previous drafts through the editor’s clock icon. That protects an editable work in progress, but it does not verify the prose, restore every external source, or publish the right version.
After Continue, Substack can ask for the intended audience, comment permissions, a free preview for paid content, tags, email and app-inbox delivery, and a future publication time. Review every option. Posts are sent by email only once; later edits update the web and app versions but do not resend the original email.
5. Use Voiceover when the author’s recording is the artifact
In a text post, Substack’s headphones menu can add a Voiceover by uploading a file or recording directly on the device. Browser recording requires Substack microphone access and an input source. Once accepted, the voiceover appears at the top of the post and readers can play it on the web or in the app.
Substack currently allows one Voiceover per post. A voiceover can also interact with a paywall and, when a Substack podcast is configured, can be added to podcast feeds. It is recorded or uploaded audio—not editable cursor text—and it can override text-to-speech for that post.
6. Use audio embeds for sound inside the article
An audio embed is different from a Voiceover. Substack lets a creator upload or record audio from the headphones menu and place the result at the top, between text, or at the end of an ordinary post. Multiple audio embeds can appear in one post.
Substack says embedded audio cannot be downloaded by the creator or listener through that feature. It can appear below a paywall and is playable on the web or in the app; an email opens playback in a web browser. Confirm placement, access, and the final post before publishing.
7. Keep podcast transcription separate from live dictation
For an audio post, Substack can generate a transcript after the creator chooses Create → Podcast and records or uploads the supported audio file. The creator can edit the generated text, upload a transcript file, assign speakers, and decide whether the transcript is shown on the episode page.
Substack says this synchronized transcript is available on the web publication, not in the app or email. Its access follows the episode: a paid podcast transcript remains for paid subscribers, while a free post’s transcript cannot be paywalled separately. This saved-recording workflow is not a substitute for live text input in an article editor.
8. Understand reader text-to-speech at its real boundary
Substack’s reader text-to-speech plays an eligible written post in the Substack app. Current documentation says it is available on most but not all publications, applies to English posts, and may be unavailable on some older posts. Eligible creators can choose a read-aloud voice in publication settings.
Text-to-speech reads existing post text; it does not put words into the editor. A creator Voiceover takes precedence when one is attached. Keep author speech-to-text, author-recorded audio, podcast transcripts, and reader playback as four separate features.
9. Troubleshoot the text and recording paths separately
If Mac Dictation produces no text, test the same shortcut in TextEdit. If it fails there too, inspect Keyboard → Dictation, the selected language, shortcut, microphone source, network requirement, and whether Voice Control is active. If TextEdit works, return to a plain Substack article body, refocus the editor, and test one unsent sentence.
If Substack Voiceover or audio recording fails, inspect the browser’s microphone permission and input source instead. If a podcast transcript or reader text-to-speech option is missing, verify that the current content type, publication, post language, platform, access level, and feature eligibility match Substack’s documented requirements.
10. Add IraVoice only for the separate local Mac input path
IraVoice 0.7.1 is a separate option for macOS 26+ Apple-silicon Macs when speech recognition and supported formatting should run on the Mac across supported Substack, browser, email, note, and developer fields. Models download on first use; the full trial needs no activation, and paid continuation needs one Gumroad key check.
IraVoice does not sign in to Substack, read publications or drafts, choose an audience, add links or paywalls, record or upload audio, create Voiceovers or podcasts, generate episode transcripts, configure read-aloud voices, schedule posts, send newsletters, or press Post or Publish. It targets a supported focused editable field and falls back to the clipboard when direct insertion is uncertain.