Voice to text in Salesforce: dictate any field

Press a hotkey, talk, watch the field fill in

The whole interaction is one motion. Click into the Salesforce field you want to fill — a case comment, an opportunity description, a Chatter post. Hold the hotkey. Talk like a normal person. Let go. The text appears at the cursor a beat later.

The default hotkey is Ctrl+Space on Windows and Command+Option on macOS. It is global, which is the part that matters — not bound to a browser tab and not a Salesforce add-on, it works in any app where you can type. The same key dictates into a case note, then the Slack message telling your manager it is handled, then the email to the customer. One muscle memory, every box. The browser extensions that show up for this search live inside a Chrome or Edge tab; Whisper sits at the operating-system level, so it does not care which browser you opened Salesforce in.

Does Salesforce have built-in voice to text? Mostly no

Here is the honest answer, because the search results are not consistent about it. Salesforce has no native, type-anywhere dictation in desktop Lightning Experience. There is no "dictate into this field" button on a case or a lead. That does not mean zero voice — Salesforce has several voice features, just not the one most people picture when they search this phrase.

The closest thing Salesforce ever shipped to native voice note capture was the Einstein Voice Assistant. It launched in 2020 and was retired in 2021. It is gone. If you are searching for an "Einstein Voice Assistant alternative," that is why — the product was real, and then it was not.

What is live today is a different category. Salesforce's standing native voice feature is Einstein Conversation Insights, which transcribes sales calls and meetings and surfaces insights from them — who talked too long, where the deal wobbled. It is call intelligence, not field dictation; it will not put a sentence into a case comment. Service Cloud Voice transcribes live phone calls for contact-center agents — again, calls, not typing. And Agentforce, Salesforce's conversational AI agent, has voice-to-text, but as a mobile feature for talking to the agent, not a system that types your words into the focused desktop field.

So Salesforce covers recorded calls natively, and you should use those tools for that. Talking a 90-word note into the activity log on a lead is the gap. It sits on desktop Lightning, and it is where a system-wide dictation hotkey earns its keep.

The phone keyboard mic versus desktop dictation

On mobile you already have a workaround, and it is fine. Inside the Salesforce mobile app you can tap the microphone on your phone's keyboard — iOS or Android dictation — and talk into a field. The keyboard mic is a device feature, not a Salesforce one. That is why mobile feels solved and desktop does not. Windows ships Win+H voice typing and macOS ships Dictation, and both type into a browser field including Salesforce — the free OS-level baseline most people forget. They are single-platform, the cleanup is thin, and toggling them is clunkier than a held hotkey. Whisper is the desktop version of that phone keyboard mic, except it works the same on Windows and Mac, in any browser, and in every other app you touch.

How to dictate into any Salesforce field with Whisper

The setup is short.

1. Install Whisper on your Windows PC or Mac and sign in. The local pipeline is free, no card required to start.
2. Download a model once — a file on disk, about 140 MB to 3 GB depending on which you pick. This is the only step that needs the internet.
3. Open Salesforce and click into the field you want to fill — a case comment, an opportunity description, a Chatter post, the email composer.
4. Hold the hotkey — Ctrl+Space on Windows, Command+Option on Mac — and talk.
5. Let go. The transcript pastes at the cursor. Edit if you need to, hit save.

No AppExchange install, no admin approval, no per-seat add-on. Whisper does not log into Salesforce or touch your org's settings — from Salesforce's point of view, text simply appeared in a field, the same as if you had typed it. That is why it works across cases, leads, opportunities, tasks, Chatter, and the email composer with no per-object setup.

Local or cloud: which mode for CRM notes

This matters more for CRM data than for almost anything else you would dictate.

In local mode, Whisper runs completely offline. The audio never leaves the machine; the only time it needs the internet is that one-time model download. When the text is client names, deal sizes, contract terms, and the occasional candid note about a stakeholder, that is not a nice-to-have.

Here is the opinion I will stand behind: cloud-only dictation is a privacy disaster waiting to be transcribed. I once watched a team ship every utterance from a dictation prototype to a cloud API. The quarter-end bill was five figures, mostly from re-transcribing the same standup recordings four times because the retry logic was too eager (I wrote that retry logic — I have a master's degree). The CFO's read was blunt: maybe do not pay a vendor to keep copies of meetings that already have notes. CRM data has the same shape. Your pipeline does not need to live in a vendor's logs because someone wanted to talk instead of type.

Cloud mode exists too, for when you want the best accuracy or AI cleanup on top — a Whisper Pro feature using your own OpenAI key. It is the escape hatch, not the default. The entire local pipeline is free for signed-in users, with the card only asked at the Pro upgrade, never at signup. The numbers live on the Whisper pricing page. For everyday CRM notes, start local, keep the data on your desk, and pay nothing.

Clean up the dictation before it lands in the field

Raw dictation has the ums and run-on sentences of actual speech. Whisper can run an optional AI cleanup pass that tidies a transcript before it lands — fixing punctuation, dropping filler, turning a rambled thought into a clean case note. In free local mode that pass runs on your machine via Ollama; in Pro it uses your OpenAI key. "Yeah so I called him back, he wants the revised quote by Friday" becomes a line you would not be embarrassed to leave on the opportunity. It is the same trick that lets you type faster with your voice anywhere, not just in a CRM.

Whisper handles over 90 languages across both modes, reaching 99 on its multilingual variants — the English-only variants are English only. If your accounts span regions, the dictation does too.

The honest limit, and when to skip Whisper

First the caveat, because the AppExchange apps make a claim Whisper does not. Whisper pastes into the single field that has your cursor, one at a time. The case note, the lead description, the opportunity field, the Chatter post — wherever you clicked. It does not understand the Salesforce data model. It will not hear "log a call, set the next step to a demo, and update the close date" and route those three facts into three record fields. It is the simpler thing on purpose: point the cursor, talk, the text lands there. The same one-field-at-a-time model is how Whisper works in a ClickUp task description and in HubSpot too.

So here are the cases where I would rather you used something else.

• You need call transcription, not note dictation. Use Einstein Conversation Insights. It is native, it transcribes the calls, and it surfaces coaching insights Whisper cannot. Whisper does not record calls; it dictates text.
• You need auto-field-routing. If you want one spoken paragraph to populate several record fields, reach for an AppExchange app like Voice Assist, Outloud, or Rollio. They are CRM-aware in a way Whisper is not, and they ship as managed packages with admin controls.
• You only work on mobile. Your phone keyboard's mic already dictates into Salesforce mobile fields for free. Whisper is a desktop tool; it earns its place on the laptop, not the phone.

Reach for Whisper when you want desktop Lightning dictation into any field, offline so the CRM data stays on your machine, free with no card and no AppExchange install, and one hotkey for every app — not just a browser tab. Setting it up on a PC? The voice to text on Windows guide covers the hotkey and OS specifics.

The first version of Whisper got hacked together on a flight because typing meeting notes was eating evenings that were already short. Salesforce reps have the same problem, dressed in a CRM. You do not need Salesforce to grow a dictation button, and you do not need to wait for the next Einstein. Click into the field, hold the key, say the thing. Download Whisper and stop retyping the call summary.