Seeking Tips for Optimizing TalkTastic Speech-to-Text Workflow

Hey team 👋 - I've been testing out different speech-to-text systems and actually prefer TalkTastic, but I'm wondering about workflow efficiency. Currently I'm recording and transcribing, but then spending time rereading and correcting the output. It feels like I'm using about the same total time as just typing it out (I'm not super fast but consistent). Anyone have feedback on trusting the system more or tips for making it more efficient? Right now I'm not sure if it's helping streamline getting ideas out or just adding an extra step. (recorded on TalkTastic)

➕2

7 comments

Arvind
·
·
I’ve had the same experience, Zach. It’s challenging to fully trust the app to produce perfect output on the first try. I use tools like Talktastic when I need to brain dump my thoughts, even if they’re not in perfect order or completely coherent. Take long emails or post-meeting follow-ups for example - you get a decent first draft, but there’s still cleanup work needed. Really hoping this improves with time. 🤔
Matt Mireles
·
·
We're working diligently to improve both transcription accuracy and the UI. I know we throw a lot at you right now - the interface definitely needs work. I've made some significant internal company changes in the past week to accelerate our progress. I completely hear your concerns and agree we need to step up our game on accuracy, usability, and most importantly, saving you time - that's the whole point of what we're building here. Really appreciate this feedback.
🙌1
Matt Mireles
·
·
I very much hear you
Zach
·
·
Love it and thanks for the feedback. The issue is more when I'm sending a message to someone. When I'm using it to prompt an AI, less concerned about the final perfect output. Keep it up! 🚀
Matt Mireles
·
·
Got it, that totally makes sense.
Zach
·
·
On this note, would be great to be able to leverage context and set some custom rules based on the context. For example, in cursor, i prefer the rewrite, but in slack, I like the transcript...etc.
Matt Mireles
·
·
Got it, that makes sense Zach. I'm working on our product roadmap right now, and this is actually a feature we're prioritizing before launching paid subscriptions. Instead of custom rules, we're taking a personalized, ML-based approach where the system predicts what you want based on your feedback and adapts to you. I definitely agree that what you want vary based on context - what I'm doing, what I'm saying - and right now the "choose for me" feature is pretty basic. That definitely needs to get smarter!
✅1