Dictate in any app · Notes
Obsidian has no built-in dictation, and its local-first users care where the audio goes. With Typally you hold a hotkey, speak, and the text lands in Obsidian exactly where your cursor was.
Yes, we build Typally. That is also why the section on Obsidian's built-in option is honest: people who switch after trying both are the ones who stay.Nothing native. Community plugins exist and some are good, but they mean trusting a third-party plugin with a microphone, and the Obsidian crowd is unusually careful about that.
Obsidian users write a lot and value keeping their vault on their own disk. Any dictation tool that quietly syncs notes to a server is a bad fit for the reason people chose Obsidian in the first place.
In practice people use it for: Daily notes, literature notes, zettelkasten entries, project logs, and the long-form thinking that ends up linked into a graph.
There is no Obsidian extension or add-in involved — Typally behaves like a very fast keyboard:
Install, sign in, and it waits in the tray.
Click first. Typally types into Obsidian wherever the caret already is, with no idea what that field is for.
Nothing truncates part-way through.
Markdown characters spoken aloud rarely land the way you expect. Dictate plain prose and add formatting afterwards — it is faster than correcting mangled syntax.
Typally types into your vault's editor without installing a plugin, which suits a community that is careful about what gets microphone access. Your notes stay in local markdown files on your disk; the audio is what goes to a transcription service. Dictate prose and type the double-bracket links by hand.
A PhD student finishes a paper and dictates a literature note straight into her vault: what the paper argues, where it disagrees with the two she read last week, and what she wants to check. The note runs three hundred words instead of the two-line stub she would have typed, and the wiki-links she adds by hand afterwards take seconds.
Dictate the note body and type the wiki-links. Link syntax is fiddly to say out loud, and the double brackets are quicker with two keystrokes than with a spoken instruction.
Markdown characters spoken aloud rarely land the way you expect. Dictate plain prose and add formatting afterwards — it is faster than correcting mangled syntax.
Not natively. Community plugins add it, and system-level dictation tools work without installing anything into the vault.
The note stays in your local vault. The audio, however, is sent to a speech-to-text service by any cloud-based tool, Typally included — the transcript is what stays on your machine.
Yes, if the tool supports it. Whisper-based transcription covers a wide range of languages and Typally detects which one you are speaking.
You can try this on the free plan — 100 minutes a month, no card, no trial countdown.
Also dictate in: Notion · VS Code · Google Docs · see all apps