GoKawiilI'm constantly seeing ads for Wispr Flow, an AI-powered transcription tool. The pitch—that you'll be able to write faster by talking out loud instead of typing—is compelling, especially if you're a slow typist. The marketing promises you'll be able to "write at the speed of thought, 4x faster than your keyboard."
I already type faster than I can think. (Fast typist, or slow thinker? You decide.) But Wispr Flow's core promise isn't just transcription—it's post-processing. The tool uses two steps. First, modern AI transcription tools turn your voice into text; second, a large language model (LLM) removes filler words and formats your words into complete sentences and paragraphs. The idea is that you can talk out your ideas and watch them turn into properly formatted text. This works inside any text box on your computer or phone.
I've tested this a few times and have to admit the results are pretty good. Apple's dictation feature, free on all its devices, works well enough—so does Google's Assistant Voice Typing on Pixel phones (which is getting another AI upgrade soon). But there's real value in software that removes filler words and formats everything into paragraphs. And Wispr Flow is sleekly designed, guiding you through the setup process with snappy graphics.
Courtesy of Wispr Flow
So what's the catch? Price. WisprFlow costs $144 per year (billed annually) or $15 a month after an extremely limited free trial. But the technology Wispr Flow is built around—AI-based transcription and LLMs—is widely available. On the speech-to-text side, Nvidia's Canary and OpenAI's Whisper are both open source, meaning they're completely free to run on your own device. And most AI enthusiasts are already paying for OpenAI, Claude, or Google's Gemini, any of which can handle the post-processing part of Wispr Flow. So can free local tools like Ollama, Google Recorder, or Apple Intelligence.
With all this in mind, I've been wondering: Is there a good, free platform-agnostic alternative to Wispr Flow? I tried out several applications—here's what I found.
Spokenly, the Best Free Alternative
If you want to get the benefits of Wispr Flow without a subscription quickly, you could do worse than Spokenly, available on both macOS and Windows. It's not open source, but it is free to download and does not require an account to use. There's a Pro plan that costs $10 a month or $100 a year. The paid plan is only necessary if you're using Spokenly's cloud models. You can opt to use a local model instead, which is free. Alternatively, if you're already paying for a service like OpenAI or Groq, you can add your API key and use that for transcribing—that's free with Spokenly.
Spokenly offers optional post-transcription formatting. You can also choose a different LLM provider for the post-transcription formatting of text. As a Mac user, I opted to use Apple Intelligence—it's totally free and worked really well in my tests. But it supports OpenAI, Anthropic, and Groq, plus a few other LLM providers. The application also allows you to write as many custom prompts for post-transcription processing as you like, each with its own keyboard shortcut.
One of my favorite things is that Spokenly can work entirely offline. If you use a local model for transcription and a local model like Apple Intelligence for the post-transcription formatting, the entire thing works without any data leaving your computer. That's nice from a privacy perspective, and from a functionality standpoint, the feature will work even when your internet is shaky.