Skip to main content

4. WisprFlow

Note: The video covers material not in the guide below — please watch in full.

Action Step

Complete this before moving on.

Go to WisprFlow settings and set up hands-free mode with a hotkey so you can dictate without holding a key. Then take some time to try WisprFlow out in a bunch of different apps — Claude Code, Slack, Google Docs, email — and get the feel of it. It's going to be one of the most important tools in your stack.

Comment in Slack

Post your answer in your onboarding channel.

What was your biggest takeaway(s) from this training?


Training Guide

You installed WisprFlow in the last training. Now let's talk about why it's one of the most important tools in your stack.

Here's the problem: when you work with an AI agent, you're having conversations. Long ones. You're explaining what you want, giving context, describing changes. If you're typing all of that, you're slow — and you lose your train of thought halfway through formatting a sentence.

WisprFlow fixes that. You talk, it types. Not robotic dictation — it actually understands how you speak and formats it the way you'd write it. Punctuation, paragraphs, tone — all handled.


How to Use It

The shortcut: Press FN + Space (Mac) or Ctrl + Win + Space (Windows) to start dictating. Talk naturally. When you're done, it converts your speech into clean, formatted text — right where your cursor is.

This works everywhere. Claude Code, Slack, Google Docs, email — anywhere you can type, you can dictate instead.


Hands-Free Mode

By default, you hold down the shortcut while you talk. But for longer dictation — like giving your AI agent a detailed prompt — you want hands-free mode.

How to activate it:

  • Mac: Press FN + Space (same shortcut, but don't hold it — just tap). Or double-tap your dictation hotkey
  • Windows: Press Ctrl + Win + Space

WisprFlow keeps listening until you stop it. Press FN once (Mac) or Ctrl once (Windows) to stop, or click the red stop icon in the Flow bar.

This is the mode you'll use most. Tap the shortcut, talk for 30 seconds, stop. Your entire prompt is written out, formatted, ready to send.


WisprFlow on Your Phone

WisprFlow also works on iPhone and Android. Same idea — talk, it types.

This is useful for capturing ideas when you're away from your computer. Think of something on a walk, in a meeting, in the car — open the app, dictate it.


Talk to It Like a Person

Here's something you need to hear early: most people tense up when they sit down in front of an AI agent. They start writing these short, careful, Google-style queries. "Create a project plan." "Summarize this document." One line. Hit enter. Wait.

That's not how this works.

An AI agent isn't a search engine. You don't need to compress your thought into five words. You need to give it more — tell it what you're working on, what the context is, what you've already tried, WHY you're trying to do this, what the output should look like. The more relevant context you give it, the better it performs.

That said — don't dump everything at it either. The AI has limited memory (we'll talk more about what that means when we talk about context window). Give it signal, not noise. We'll cover how that works in upcoming trainings. For now, just aim for the sweet spot: talk to it like you're explaining something to a coworker — give them what they need to help you, not your entire life story.

And that's exactly why WisprFlow matters. Because nobody's going to type three paragraphs of context. But you'll absolutely say three paragraphs — it takes 20 seconds.

This is a habit you'll need to retrain. You've spent years being efficient with search — short keywords, specific phrases. With an AI agent, that instinct works against you.

You'll see me do this throughout the trainings. I'll use WisprFlow to talk to Claude Code, and you'll notice the prompts are long, conversational, and specific. That's not me being wordy — that's how you get great output.


Voice and Typing — It's a Mix

WisprFlow isn't here to replace typing. It's another tool in the belt.

You'll use voice when you're explaining something, giving context, brainstorming, or rattling off a long prompt. It's great for getting ideas out fast without overthinking the wording.

You'll still type when you want precision — when you're editing a specific line, pasting a file path, or crafting something exact. Sometimes writing by hand helps you think more carefully, and that matters too.

Most people end up with a natural mix. Voice for the bulk of it, typing for the details. Don't force yourself into one or the other — just notice which one fits the moment.


How This Fits Into the Workflow

At LeanScale, the core workflow is: voice → text → AI

You think of what you want. You say it out loud. WisprFlow turns it into text. You send it to Claude Code. The agent does the work.

This is faster than typing for most prompts, and it keeps you in flow. You're not editing sentences in your head before typing them — you're just talking. The AI doesn't care about perfect grammar. It cares about clear intent.

(Next up — we're going to talk about the big idea: what makes this fundamentally different from ChatGPT)