How I Run Speech to Text on Fedora Linux for Free

I switched to Fedora as my daily driver last year and generally have enjoyed it. I've also been doing a lot of vibe engineering - using AI to speed up my dev workflows.

As I've streamlined my vibe engineering workflows, I've found that dictation (speech to text) is useful for quickly providing a lot of context to AIs. AIs do better with more context so it's useful to have a quick way to do this and dictation is typically a LOT faster than typing - Fast typers can type 120 - 150 wpm while I'm sitting around 100 and most below 60 while most people can speak 120-160 wpm easily with dictation.

Mac has pretty good dictation built in and Mac / Windows both have access to commercial solutions like Whisperflow but I haven't found any off the shelf software I was super happy with on Fedora.

So in this post I wanted to share my workaround so you can use it if you're in a similar position.

Speech to Text on Fedora

My solution is pretty dumb, requires an internet connection, and requires some copy pasting but it still saves me time when trying to context dump paragraphs for an AI.

My solution: Google Docs and their built-in dictation

Why this works:

• Well supported across OS and distros - Google Chrome is very well supported with active development and GDocs runs in the browser
• Free - If there are limits to dictation, I haven't seen / hit them
• Relatively fast

Downsides:

• Requires an internet connection (though most SOTA AIs also require this so don't mind too much)
• Not the most accurate (but most AIs can figure out what you meant even with misspellings / grammar issues - and you can always review their plan / answer followups to make sure)
• Still need to copy paste (but still saves time for large context dumps)

I have a bookmarked Google Doc called "Speech to Text Doc" that I use for all my adhoc speech to text. Here's what it looks like:

# Speech to Text doc

Goal: Provide a simple dumping ground for speech to text using GDocs. Helpful on systems that don’t have great integrated stuff.

To use: 

* Put cursor in playground area  
* Hit CTRL+SHIFT+S to start mic  
* Start talking  
* Cut and paste your words to wherever you need them

## Sandbox: (Cursor below here \- will be deleted later)

Next

So that's how I'm doing speech to text. I'd love to have a reliable, local-first version that was fast, cheap, and could run anywhere with output appearing at my cursor but haven't found one yet.

I'm not quite to the point where I'm ready to build my own version as that seems like it will require a lot of learning ab different distros, AIs, and rendering systems but might end up being a good project for my time at Recurse Center.

If you liked this post you might also like:

• 5 AI Coding Best Practices from a Google AI Director (That Actually Work)
• Stop Vibe Coding, Start Power Coding - How To Write Quality Software Faster With Agentic AI
• How to Change Login and Lock Screen Backgrounds on Fedora 42 KDE Plasma