Guide›Voice & AI
Voice & AIBeginner8 min

Voice brainstorming

Describe your idea out loud like you're talking to a colleague. Myndlab listens, asks the right follow-up questions, and turns the conversation into a build prompt.

What is voice mode?

Voice mode lets you speak your idea rather than type it. It's especially useful when you have a clear mental picture of what you want to build but find it faster to explain verbally — or when you're not sure how to structure a written prompt.

Under the hood, Myndlab transcribes your speech, analyses the intent, and starts a structured back-and-forth to fill in any gaps. By the end of the conversation, it has assembled a precise build prompt you can review before committing.

Enable voice input

1
Open a project
Go to any existing project's Build screen, or create a new project from the dashboard.
2
Click the microphone icon
The mic icon sits at the right end of the prompt input bar. On first use, your browser will ask for microphone permission — click Allow.
3
Wait for the wave indicator
A small animated waveform confirms that Myndlab is listening. Start speaking when you see it.
ℹ
Note.Voice mode works in Chrome, Edge, Arc, and most Chromium-based browsers. Firefox support is partial — some speech APIs may not be available.

How the conversation works

Voice brainstorming is a short multi-turn conversation, not a single monologue. Here's what a typical session looks like:

1
You describe the idea
Talk through what the app does and who it's for. You don't need to be precise — broad strokes are fine at this point. Aim for 30–90 seconds.
2
Myndlab asks follow-ups
Myndlab pauses and asks 2–4 clarifying questions out loud (and shows them as text). Answer each one by speaking; you can also type if you prefer.
3
Summary is assembled
After the follow-up round, Myndlab reads back a summary of what it understood. You can correct anything it got wrong before proceeding.
4
Prompt is generated
Myndlab compiles the conversation into a structured written prompt and puts it in the input box. You review it, edit if needed, then hit Build.

What Myndlab asks you

The follow-up questions are generated based on what was missing or ambiguous in your initial description. Common ones include:

  • Users and roles — single user, a small team, or public-facing with sign-up?
  • Key data — what are the main things the app stores or tracks?
  • Core actions — what's the most important thing a user does in the app?
  • Integrations — does it need to connect to Stripe, Slack, a third-party API?
  • Stack preference — any technology you specifically want or want to avoid?

You don't need to answer every question exhaustively. If you're unsure, say so — Myndlab will pick a sensible default and note it in the prompt so you can change it later.

Review before building

The generated prompt appears in the text input before any code is written. Take a moment to read it through:

1
Check the feature list
Does it capture everything you described? Anything missing or misunderstood?
2
Verify the stack
If you have a preference, confirm the right framework and database are listed.
3
Trim or expand
You can type directly in the input to adjust wording, remove a feature you're not ready for, or add a constraint you forgot to mention.

When you're satisfied, press Build →. The rest of the process is identical to a typed prompt.

Tips for better results

💡
Tip.Speak as if you're explaining the app to a new colleague — conversational, concrete, and specific. "Users can filter by status and date range" is better than "the app should have filters."
💡
Tip.If Myndlab misheard a word in the transcript, tap the text and fix it before continuing the conversation. The correction propagates into the final prompt.
ℹ
Note.Voice sessions are saved to your project history. You can revisit the transcript and prompt from Project Settings → History if you want to reference what was discussed.

Voice brainstorming is also useful mid-project, not just for the initial build. When planning a new feature you can start a fresh voice session on an existing project and the context from previous builds is included automatically.