Big Macrowhisper Update! More Automation Possibilities with Superwhisper
Context stacking for AI, chained scripts, text transformations, and lots more improvements!
Hey guys,
I’ve been spending the last couple of weeks working pretty intensely on a new Macrowhisper update, and I just released a video showing some of the new stuff. So instead of the usual recap, I thought I’d give you a quick update on what I’ve been building. I didn’t really get the chance to sit down and consume much content like I normally do, but hopefully this will still be of interest for you.
Gracias, Codex
Over the last few newsletters I mentioned that I had the chance to play around with Codex for a while. First through the free access that was available as part of their latest release, and later through a trial opportunity that popped up when I had reached the limit. That ended up being the perfect excuse to revisit a lot of ideas I had been considering for Macrowhisper but had never attempted to implement. In big part because I didn’t have the time and because most development until now has been via the API (which gets pretty expensive). There’s also the other side of things where I just didn’t know if I’d be able to pull this off, even with AI’s help 😅
So for the last couple of weeks I decided to just push forward with this and see how far I could get. Now, I talk about this tool all the time and I am assuming that if you follow here you already know about this, but Macrowhisper is a project I made to extend what you can do with Superwhisper. I’ve always loved the dictation side of Superwhisper, and since I’m also very into automation, Macrowhisper grew out of trying to connect those two worlds. The whole thing would have never been possible without AI, and this update is the result of even more experimentation.
Lots of minor things
There’s been a lot of improvements on chained actions, script execution, and the clipboard system. There’s several new features (like transformations, smart insertion, and new CLI flags) but the biggest changes have been behind the scenes. Some parts of the app were refactored completely. For new users the experience probably won’t be drastically different (hopefully the whole thing will make a bit more sense), but for people who already use Macrowhisper there are quite a few improvements in how things are organized and how the configuration works. Now the rules are more consistent and easier to read at a glance. All the details are up on the release notes.
One thing that’s been clear to me is that Macrowhisper can look a little intimidating if you’re not very technical. So alongside this update I rewrote the documentation, reorganized things, and added more examples. My hope is that this makes the project easier to explore even if you’re not someone who normally likes digging through CLI apps or config files.
Thanks guys!
This past week was pretty intense working on this update, so I didn’t get to do much reading, watching, or experimenting with other tools like I normally do. Next week the substack should be back to the usual format.
Thanks again for following along!
If you find this useful, I would be incredibly grateful if you could support me by buying me a coffee at THIS LINK. Your generosity would mean the world to me.
If you liked this you may also enjoy some content I have up on my YT Channel! I don’t hang around social media a lot, but when I do I’m on IG or Twitter. You can also check out some of my online classes, listen to my music, or in case you haven’t already, subscribe to my weekly newsletter. Thank you for reading!


This might be better left as a comment in the GitHub. But I was wondering about how the URL actions work. I've been trying to trigger a URL action while being in the website without any text within my actual dictation. For example sometimes I'll copy a text and open up I'll copy a text within the answer for my response from some AI chat that I'm using, and then quickly trigger and stop the dictation, hoping that my action for asking that question triggers. But it looks like for it to trigger, I have to have some sort of dictation recorded.
Could there be a way that the URL actions trigger just from a quick start and stop of a dictation while I'm on the website?
This leads me to another question though. How does Macro Whisper know that I'm on a specific website when I do a voice-only dictation with context application off? I've been checking the JSON files of some of the dictations I'm doing while on the app, and there's no actual data within the JSON file to suggest that I'm on a specific website, but my action will still trigger correctly (as long as I have an actual dictation).
Sorry for so many questions. Two more thoughts.
1. Could there be a way that we could make the CLI command persist through the dictation? So that the clipboard is constantly being updated after I trigger the script. This would help when I am planning a very long dictation, but I want to just trigger the clipboard stacking feature early on. This would make my below question applicable, since I I just realized now that the command, if I run it at the end of the dictation, would actually not allow me to know which clipboard I need to get rid of anyways
2. Second, is there a way to send a stacked clipboard to the clipboard before I start? I just tested triggering the CLI command before re-processing, and it works well if my buffer is long enough, setting up my stacked clipboard, but then once I hit re-process, it seems to clear my clipboard. I was running some example tests and I don't really understand what "clipboard" a re-processed dictation uses