Question · Voice and AI
Is there a free speech to text AI?
Yes. You can get free speech to text in three ways: the dictation built into your phone or computer, open-source Whisper apps that run on your own machine, and free plans or allowances from dictation apps and cloud services. Each one has a catch, such as a word cap, a monthly hour allowance, a need for an internet connection or the setup time of a local model.
Updated October 17, 20264 min read8 cited facts, each traced to its source
Yes, and there are several. The free options differ in what they cost you: a word cap, a monthly allowance of audio hours, a need for an internet connection, or the setup time and memory of running a model yourself. Choose by the job you have, and check the limit before you rely on it.
Open-source Whisper is free to use and runs on your own computer
OpenAI's Whisper is a free speech model that you can run yourself. Its README says that "Whisper's code and model weights are released under the MIT License" 1, a permissive license that lets you use the software without a per-minute bill. You still need to install it, which means an app or a command-line tool and a model download first.
Several apps wrap Whisper for everyday use. Handy describes itself as a free, open-source speech-to-text application that works completely offline 2. Because the model runs on your computer, your audio does not have to travel to a server. The trade-off is setup time and a machine with enough memory. Our guide to running Whisper locally gives the steps, and whether Whisper is free covers the costs that remain.
Dictation apps offer free plans with a cap
Some dictation apps let you try them without paying. Superwhisper's free plan gives unlimited dictation using local Whisper models, with up to two modes, a custom vocabulary and text replacements 3. The cap there is on features, not on minutes.
Other apps cap the words. Aqua Voice's free plan includes 1,000 words and its Avalon transcription model 4. That is enough to test an app on a few emails, but it is a small allowance for daily dictation or a long report. Read the plan page for the cap before you start, because it can change.
Cloud speech services have free allowances, mostly for developers
Cloud services sell speech recognition by the hour, and each one gives a free allowance. These are built for software that sends audio to an interface, not for a finished dictation app, but they are useful if you are building something or testing accuracy on your own recordings.
| Service | Free allowance | What to check |
|---|---|---|
| AssemblyAI | Up to 185 hours of pre-recorded transcription 5 | Live (streaming) audio has its own allowance on the same pricing page |
| Azure AI Speech | 5 audio hours a month of standard real-time transcription 6 | Standard real-time transcription only |
| Amazon Transcribe | Up to 60 audio minutes a month 7 | Only for the first 12 months after signup |
These allowances look generous until you dictate every day. A daily dictation habit can use up a monthly allowance quickly, as the cost guide works through.
Built-in dictation comes with the device, and it depends on the network
The dictation built into your phone or computer comes with the device you already own, but it depends on the operating system and the network. On Windows, voice typing needs an internet connection, a working microphone and a cursor in a text box before it will type what you say 8. If you are offline, that built-in tool may do nothing, and an open-source app may be the better choice.
It depends on what you need the text for
- Short notes and messages: start with the built-in tool. It is already on the device, and short text rarely needs more.
- Offline or your own machine: choose an open-source Whisper app. You pay with setup time and memory.
- Trying a paid app: use its free plan first, and note the cap, because heavy use runs into it quickly.
- Building a product: use a cloud allowance, and plan for what happens after the free period ends.
- Daily, all-day dictation: a free plan will not cover it. Compare the cost of a subscription or a per-minute service against your own minutes, as the cost guide shows.
Free does not mean private. A free cloud service sends your audio to the company that runs it. Before you dictate client, medical or personal information, read whether free AI dictation is private.
What to do
- Try the built-in dictation first in the two or three apps you use most, for a few days. If it is accurate enough, you have a free answer.
- If you need offline use, install an open-source Whisper app and dictate one page to test its speed and accuracy on your computer.
- Check the cap or allowance and its reset date before you depend on a free plan.
- Read the privacy page of any free app that sends audio to a server, and see whether AI dictation apps send your text to the cloud.
- For free file transcription in the browser, the audio to text tool works on recordings you already have.
Yap, which we make, runs dictation on the device on Mac and Windows, so it falls into the on-device group above. Current plans are on our pricing page. For the wider picture of who pays for free speech recognition, see the guide to AI dictation costs, and the free dictation software guide for more free options by platform.
Evidence
Every number on this page, with where it comes from. Open a fact to see the exact words in its source.
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
Go deeper
Related questions
- Does on-device AI dictation drain your battery?
- Is free AI dictation private?
- Is on-device dictation cheaper than cloud dictation?
- Why do AI dictation apps charge a monthly subscription?
- Are AI summaries of voice recordings accurate?
- Are smart speakers still popular?
- Can AI detectors flag dictated text?
- Can AI dictation run entirely on your own computer?