As an Engineering Manager at Buffer, I write a lot — Slack messages, proposals, code, documentation, and feedback. When AI became a part of my work, I wrote even more.
It’s always taken me a long time to type, and my brain works better when I speak than write. Before, it never seemed like I was limited, but since the advent of AI, this feeling has become more pronounced. It was easier to keep the questions short than add additional detail.
When I used the voice option in ChatGPT, something clicked. This allowed me to include context and nuance in my messages that I would normally leave out. I didn’t want that app to be the only one.
In my daily work, I’ve tested out four AI-driven speech-to text tools over the past year. The speed, accuracy and features of each tool are different. Learn more about my experience using each tool. You’ll also learn how much it costs and how I integrated dictation in my workflow.
The power of Superwhisper needs tweaking
Get 15 minutes free of dictation each month
Paid plan: $8.49 per month
Since over a month, I’ve been using Superwhisper I was very happy with the tool. The tool is solid and offers a wide range of customization options. You can choose different AI models, prompts, modes, etc. for each application. So, it will apply different presets to Slack than a programming IDE.
Subscriptions were affordable for less than $10 a month and worked fairly well. However, I encountered two problems. My accent made it difficult to dictate accurately. I also had to edit a lot afterward, particularly with the technical terminology. The system was slow. I would dictate and wait while it processed, interrupting my flow.
It became clear that I didn’t need a complex tool to customize and tweak. I wanted something which worked well right out-of-the box. I wanted to see if there were other solutions.
VoiceInk is a low-commitment way of trying dictation
One-time Payment of $25
You tried? VoiceInk The next step was based upon a recommendation from a teammate. The cost is $25 and the software is open source, something I’m proud to support.
This is a clever user interface that integrates into macOS. The processing is also local and takes place on the device itself, which makes it a great privacy tool. The customization features are great and it was faster than Superwhisper.
The accuracy was not quite right for my requirements, so I continued to look. This is an excellent choice if you are new to using dictation software and would like to try it out without having to commit to a monthly subscription. It’s low-risk and could be a good option for those who use it occasionally or have different accuracy requirements than me. VoiceInk is used by many people at Buffer.
Wispr: The fastest, most feature-rich and refined flow.
Plan free: Up to 2,000 Words per Week
Paid plan: $15 per month
Wispr Flow What tool am I currently using?
Although it’s more expensive, the accuracy is significantly better. It’s accurate and saves the words that I have corrected.
Speed is the main difference. This technology processes audio as you speak, allowing the transcription to appear almost immediately. It has transformed my workflow, allowing me to stay in flow much more often and not have to switch contexts as frequently.
Many other tools struggle to handle multiple languages. Wispr Flow is able to handle switching between English and Ukrainian, my native language.
The cost of $15/month is the most expensive option, but it has been well worth it in terms of performance and reliability for professionals who use it every day.
Willow Voice: High quality with less features
Plan free: Up to 2,000 Words per Week
Paid plan: $15 per month
Willow Voice The price is similar to that of WisprFlow, and it has many benefits as well – the software runs fast and with good accuracy. In my use case, Wispr was a little slower and less reliable, but by not a large margin.
The company seems to be in its early stages, as some features were not yet available. Although I will keep an eye out for it, I decided to go with Wispr Flow because it had a slight edge in the areas that were important to me.
| You can also use the following tool: | Buy Tickets | Platform | The speed of the vehicle | Accuracy |
|---|---|---|---|---|
| Superwhisper | $8/mo | MacOS + iOS | Medium * | Medium * |
| Voice Ink | $25 one-time | MacOS + iOS | Medium * | Medium * |
| Wispr Flow | $15/mo | MacOS + iOS | ||
| Windows | ||||
| Android | Fast | High Quality | ||
| Willow Voice | $15/mo | MacOS + iOS | Fast | High Quality |
✏️ Noteworthy:The speed will vary depending on the AI model that you use for local tools, such as Superwhisper. The larger models are more accurate, but they take longer to process.
How I’m using dictation in practice
I quickly found that dictation was a part of every day work. In my daily work, I’ve seen it manifest in the following ways:
AI prompting
Most of the time, I no longer type AI prompts in chats or when using AI agents to code. The biggest time-saver for me was this.
The AI has also improved as my prompts have become more precise and specific. For convenience’s sake, I usually keep typing brief. In dictating, it is natural for me to provide more explanation and context.
Note-taking and brain dumps
This is usually something I think about quickly and then transcribe later. These could include project ideas, thoughts after a meeting, or problems.
This is not a time to produce polished prose – I am simply capturing my thoughts into a form I can later revisit. This is possible because dictation moves faster than typing.
Instant messages and replies
I sometimes dictate shorter messages*,* too, though this varies. It’s especially useful for long explanations, where context is needed. Even if you have a very short message or a long one with many formatting options and links, typing will still be faster.
Custom workflows
I’ve developed some workflows specific to dictation. Raycast has a number of custom AI commands that can be used to process and format dictation in various ways.
Interviewing is one use I have found to be particularly useful. Buffer is on a hiring spree, so I do a great deal of interviews. An interview has resulted in an AI command. Granola This is where I transcribe and write down my ideas and impressions. The process has been streamlined significantly.
Considerations and recommendations
Throughout this experience, I have learned many things. I’ve learned a lot during this process.
- It turns out that speed has more of an impact than I thought. The few seconds you spend waiting can add up and disrupt your workflow in unexpected ways.
- I process most things through talking. I find that dictation tools allow me to think more quickly and come up with the best words for a problem. This could be a game changer for you if you are someone who likes to think out loud.
- The goal is not to replace typing completely. Different tools are needed for different contexts. Dictation is most helpful for AI interactions. It captures thoughts and ideas more quickly. And it allows me to give longer explanations. When it comes to quick replies, formatting text or collaboration editing, typing is more efficient.
- VoiceInk offers a low-risk option to get you started with dictation. Wispr flow is worth it if you use it frequently and find the limitations frustrating.
You can reach me on Twitter or Facebook if you’d like to share your experience with these tips and tools. Please feel free to contact me on LinkedIn.
P.S. This article was dictated in large part😉
