Speech recognition software: what to choose

Speech recognition software turns your spoken words into text or commands, and on Windows and Mac the first step in choosing one is knowing which of its three jobs you need.

Free for 2,200 words a week, with paid plans for more. See the plans

Facts checked on 6 November 2026 against each maker’s own website.

Shown with Advanced formatting switched on. Every plan has it.

What speech recognition software does

Speech recognition software listens through your microphone and turns what you say into something the computer can use. You may also see it called automatic speech recognition (ASR) software, or voice recognition software. It does one of three jobs:

  1. Dictation. You talk and the words appear as text in an email, document or chat.
  2. Voice control. You say commands to open programs, click buttons or move around the screen.
  3. File transcription. You give it a recording, such as an interview or a meeting, and get the text back.

Know which job you want before you compare tools. Most people searching for speech recognition software want the first: dictation.

Speech recognition built into your computer

You may already have something that works.

  • Windows voice typing. Press the Windows key + H with your cursor in a text box. Microsoft’s page says you need to be connected to the internet, and that Windows 11 has a setting for automatic punctuation. Our Windows voice typing guide has the steps.
  • Windows Voice Access. On Windows 11, version 22H2 or later, Voice Access lets you control the PC and write text by voice. Microsoft says it uses on-device speech recognition and works without the internet once set up. It replaced the older Windows Speech Recognition on those versions in September 2024.
  • Apple Dictation on the Mac. Switch it on in System Settings, under Keyboard. Apple’s page says you can dictate text of any length, that it stops by itself after 30 seconds without speech, and that it adds commas, full stops and question marks for you in supported languages.

These come with your computer and are a fair first test. Their limits: each works only on its own system, and Windows voice typing needs the internet.

Which tools do which job

As described on each maker’s own page on 6 November 2026. We have only tested Talkativ.

Job Built into Windows or Mac Programs and apps
Dictation (typing what you say) Windows voice typing; Apple Dictation Dragon Professional v16 (Windows), Wispr Flow, Superwhisper, Spokenly, Handy (free, open source), Talkativ
Voice control of the computer Voice Access (Windows 11) Not compared here
Transcribing audio and video files Not covered on the pages we read Superwhisper, MacWhisper (Mac only). Talkativ does not do this
Works without the internet Voice Access, after set-up Superwhisper, Spokenly and Handy, with local models

Dragon is the long-standing name for professional dictation on Windows; Nuance’s page shows Dragon Professional v16 for Windows 11 and 10, with no price (“Contact us”). For its current range and the alternatives, read Dragon NaturallySpeaking alternatives. Prices and free plans for the dictation apps are side by side in our buyer’s guide to voice to text apps.

Sources: Microsoft on voice typing, Microsoft on Voice Access, Apple on Dictation, Dragon Professional, Superwhisper, Spokenly, Handy on GitHub, MacWhisper, Wispr Flow.

How to choose

  1. Pick the job first. Dictating, controlling the computer and transcribing files are different jobs, and few tools do all three.
  2. Match your computer. A Mac-only tool will not help on Windows.
  3. Decide about the internet. If you need to work offline, rule out the online-only tools.
  4. Test with your own voice. Read the same two paragraphs, including your own names, into each tool. Look at the punctuation and the mistakes.
  5. Check the price after the free part. Compare weekly limits and yearly prices.

The full checklist is in the buyer’s guide. If the term “AI dictation” is new, this explainer covers how it differs from the older kind.

Where Talkativ fits

Talkativ is an AI dictation app for Mac and Windows, so it does the first job only. Click into any text box, press Ctrl + Space on Windows, ⌥ Option + Space on a Mac, speak, and press again. Your words appear at the cursor with punctuation and capital letters. You can add names to a Vocabulary list. Free is 2,200 words a week; Pro is $7 a month or $56 a year.

It is not a voice-control tool, so it will not open programs or click for you. It does not transcribe audio files, has no phone app, and does not work offline. It is press to start and press to stop, and the shortcuts are fixed. Read about Talkativ for Windows or Talkativ for Mac, or go to the home page.

Questions

The short answers. More questions

Most people use the two terms for the same kind of software, so you will see both in searches. Strictly, speech recognition works out what you said, and voice recognition can also mean working out who is speaking. Dictation tools care about the first.

Yes. Windows has voice typing and, on Windows 11, Voice Access, and the Mac has Dictation. Handy is free and open source (its GitHub page lists an MIT licence) and runs offline. Some apps also offer free plans with a word limit.

Some older tools offered it: Windows Speech Recognition, for example, has a "Train your computer to better understand you" option. Talkativ needs no voice training; you install it and talk. It has a Vocabulary list where you add names and terms so they are spelt right.

Some can. Microsoft says Voice Access works without the internet after set-up, and apps such as Superwhisper, Spokenly and Handy say they run on your computer. Windows voice typing needs the internet, and so does Talkativ.

Speak. It’s written.

Free for 2,200 words a week, on Mac and Windows.