Which open-source dictation apps are developers following?

The open-source dictation app that developers follow most is cjpais/Handy, with 34,130 stars ¹. Next come altic-dev/FluidVoice with 12,287 stars ², OpenWhispr/openwhispr with 9,573 stars ³ and Open-Less/openless with 3,893 stars ⁴.

These are all desktop tools for voice typing, which is a newer group than the research engines they build on. Our dataset does not tag a repository as a dictation app, so this page selects repositories by name and purpose. The dataset has 124 data points ⁵, and it also holds repositories under the dictation topic.

For every row, see the full dataset.

Newer dictation apps are close behind the research engines

Handy is built in Rust under the MIT license. The dataset shows a push on the day of the latest reading, so it is actively maintained. Its 34,130 stars ¹ put it ahead of long-running engines such as SYSTRAN/faster-whisper, which has 25,969 stars ⁶.

Several other dictation-style apps appear, each from a different stack:

• FluidVoice is written in Swift under GPL-3.0, so it targets Apple platforms.
• OpenWhispr is written in JavaScript under the MIT license.
• OpenLess is written in Rust under AGPL-3.0. It is among the newest rows in the dataset.

Here is the table filtered to the Swift rows, which tend to be Mac and iPhone projects.

Many of the apps sit on top of Whisper-family engines

Open-source dictation apps rarely build their own recognizer. They wrap an engine and add a hotkey, a microphone capture step and a way to paste text into the window you are using. That is why the engines still lead the list: ggml-org/whisper.cpp has 54,482 stars ¹³. Libraries for real-time use are also followed, such as KoljaB/RealtimeSTT with 10,194 stars ¹⁴.

If you want an app rather than a library, check the license and the language in the table. A GPL or AGPL license matters if you plan to include the code in your own product.

Showing 8 of 48 rows. See every row in the dataset.

Apps for transcribing recordings, as opposed to typing as you speak, are a separate group. chidiwilliams/buzz has 22,078 stars ¹⁷ and thewh1teagle/vibe has 7,823 stars ²⁰. Both handle files and recordings, so they suit meetings and interviews more than live voice typing.

Stars say what developers follow, not what works best for you

Star counts are a record of attention. Keep these limits in mind:

• Whole-history counts. Stars rarely fall, so they reflect a project's whole history. A repository that has been quiet for months can still rank high.
• Launch spikes. A launch on a news site adds thousands of stars in a day, so a young project can jump past established ones.
• Self-assigned topics. Repository owners choose their own topics, so a repository tagged dictation may do something else.
• Followers are not users. Developers who star a project may never install it, and people who use it may never visit GitHub.
• No accuracy data. The numbers say nothing about word error rate, speed or language coverage.

Use the last push date in each row as a quick health check. A project pushed in the last few weeks is being worked on. One with an old last push date may not get fixes for new operating system versions.

The first chart shows the most starred projects across the whole dataset.

The second shows how stars have changed over time for the tracked repositories.

What this means if you want to dictate

An open-source dictation app is a good fit if you are comfortable installing software from GitHub, picking a model and updating it yourself. You get control over where your audio goes, since many of these tools run on your own computer. You give up polish, support and a guarantee that the project continues.

If you only need occasional dictation, start with what your computer already includes. Free online dictation runs in your browser with no install. If you want an app that works on your device without setup, Yap, which we make, offers on-device dictation for Mac and Windows. See download to try it.

To compare finished dictation products side by side, see the market overview. For background on how the engines work, read our overview of speech recognition and the guide to open-source speech recognition. For the broader ranking of engines, see the question on the most popular open-source speech recognition projects.