Wispr Flow
What it is
Dictation that works in any application: you speak, and text appears wherever the cursor is. It can also transcribe meetings, and it offers an MCP connection, the standard assistants use to reach other software, so what you dictated or discussed is available to whatever does the next step.
What it’s good at
Getting thought into text faster than typing, which pays most on briefs, notes and instructions, where the thinking is the work and the typing is the drag. Meeting transcription when you want it, so a call can produce a record without anyone taking notes. And the connection: an assistant can read what came out and turn it into drafts, tasks or a summary, which is the part that decides whether capturing any of it was worth doing.
What it costs to own
Dictated text arrives rough, so the saving is the first draft and someone still owns the edit. Transcription is a feature you turn on for a call rather than a default. And a captured note is worth nothing until something happens to it; the value sits downstream, in whatever the connection hands the text to.
When it’s a poor fit
It fits poorly for anyone who thinks better by writing than by talking. Dictation moves the bottleneck, and for some people the bottleneck was never typing. And some conversations do not need a transcript, in which case the transcription simply stays off.
What it sits beside
It replaces typing the first draft and the colleague stuck writing the minutes. Downstream it belongs beside whatever holds the work: a task system, a CRM, a document store. If the meeting platform’s own recorder runs at the same time, two tools on one call means two archives of the same conversation.
My experience
I dictate more than I type, and it is the connection to the assistant that makes me keep it.