Dictate into any field and drive the app by speaking, so your hands can stay on your trading platform while a trade is still live. Voice mode is off until you turn it on, and everything it needs is in one place.
Voice is configured under Settings → Capture → Audio setup, not under a tab called Voice. Audio input and screen capture share a page because they are both about the hardware on your desk. The section holds four blocks, in this order:
No questions on this page match that. Search all 15 areas
Settings → Capture → Audio setup → Voice mode, tick Voice mode (always on). After that the hotkey — ⇧⌘V on a Mac by default — toggles it without opening Settings. You need a Groq or OpenAI key first; see the next question.
A supported browser, a microphone, and an API key — Groq or OpenAI. A Groq key is free, and that is the one to get. Turn voice on without a key and the app says so: Voice needs a Whisper key (Groq or OpenAI) — add one in Settings → AI.
Toggle listens automatically whenever a field is focused. Push-to-talk listens only while you hold a key. Toggle is the default; push-to-talk is more deliberate and avoids silence-detection problems entirely.
Yes, both of them. Press Rebind beside either key, then press the combination you want.
Around twenty-five, grouped below by what they do — filling in a form, starting a routine, going to a page, asking what is available, repeating yourself, and correcting what voice just did. Anything that is not a command becomes dictation into whatever field you have focused.
Read the full guideA good many, and the answer is spoken back to you. "What's my P&L today", "win rate this week", "how many losers today", "average winner", "P&L on NQ this week" and "show me my last NQ loser" all work, from any page. They need no AI key — the answer is arithmetic on trades you have already logged.
Read the full guideYes — several pages add their own, on top of the global set rather than replacing it. Say "what can I say" on any page and it reads back that page's commands first.
Read the full guideIt depends on how much you said. A phrase carrying a direction or an entry price is written straight away — no review step. A bare "new trade" opens the + Trade form instead, and you check it and press the button yourself.
Read the full guideVoice mode → Microphone. It defaults to your system default; pick a specific input if the browser keeps grabbing the wrong one.
Press ▶ Test mic under the microphone picker and talk. The bar should turn green and push past the marked line.
Two opt-in confirmations under Command feedback — a short chime when a spoken command matches, and an operating-system notification when the tab is in the background.
Reading a short script so the app learns your voice and responds only to you. Not required to dictate, but it is what stops a livestream, a podcast or someone else in the room from driving your journal.
Speaker recognition → Set up my voice, wait for the model to download, then read the ten lines out loud. You need at least five clips and eight seconds; Finish unlocks when you have them.
In order: check the level meter, check you are on the right microphone, lower strictness a little, and re-enroll if the card says your voiceprint is from an older model.
How close a match to your voiceprint the app demands before it acts. Higher rejects more. It runs from 0 to 1 and starts at 0.70.
On by default. The app quietly refines your voiceprint from clips it is confident are you, so recognition tracks changes in your voice, room and microphone without you re-enrolling.
Yes. Your voiceprint is stored on your account, so enrolling once covers every browser and machine you sign into.
Press Re-enroll on the "Your voice" card and read the script again. It replaces your old voiceprint.
Two ways. Type the term into Personal vocabulary so transcription expects it — and when you correct a value the app set, it remembers that correction on its own.
Vocabulary is a list you type in advance. Corrections are pairs the app learns from what you fix. Vocabulary biases transcription; corrections also rewrite a specific mishearing.
Your own trigger phrases that fire one or more actions in sequence. Add them under Voice macros: a name, the phrases that trigger it, and the actions to run.
Exactly what was heard and what happened to it, for the last hundred or so voice events. Reach it from Open voice diagnostics → at the bottom of the Voice mode block.
