Voicepal: your AI Ghostwriter

Currently, I've been using VoicePal for the last three months. I originally discovered it through one of Ali Abdaal's videos, and I thought it would be an interesting way to approach making scripts and journal entries. Overall, I have very much enjoyed the experience. It’s been incredibly helpful for figuring out my script templates and establishing a routine. For my weekly scripts, I go through what I want to talk about and identify my focus points. The tool really helps by asking the correct and necessary questions to help me understand exactly what I need to think about when writing. It has truly elevated my scriptwriting process. Another useful feature has been the journal entries. I ended up creating a custom "Journal Buddy" prompt. I really like that I only needed to enter that prompt once; now, I just select Journal Buddy every day, and it asks me the same base questions followed by the specific follow-up questions I’ve instructed it to ask. It does a great job with those follow-ups, but it also knows when to move on once it has gathered enough information. If there is still something I want to touch on, it's very easy for me to say that I’d like to talk more about a specific point. There's three reasons for my ratings. The voice options, the bugs, and lastly the cost. 1. Voice Options I will say that the voice options are fairly unique to VoicePal, and I do enjoy the concept. I like the idea that when I use my "Script Writing Buddy," the voice can be completely different—like I’m talking to a specific script ghostwriter—compared to the voice for my journal entries. The problem I’ve run into is that I don’t like many of the other voices. Sorry, Ali, but your voice is actually the worst out of all of them. To me, it doesn't come off as friendly; it seems cold and quick. For some reason, it doesn't seem to ask many follow-up questions, and it feels uninterested in the conversation. These are just my feelings based on a couple of tests, but those tests made me not want to continue using that voice or the others provided. They don't seem friendly enough for me. If that tone is intentional, then I hope you have more friendly voices coming down the pipeline. I want an assistant that sounds interested in the conversation; otherwise, I feel like I shouldn't be talking as much, which makes me want to just get it over with rather than really enveloping myself in the creative process. Honestly, I hope more apps adopt this kind of variety—specifically Claude. Right now, I use Claude’s Australian voice because that sounds the most friendly to me. I hope you add more voices in the future or update the current ones to make them sound a bit warmer. 2. Technical Bugs and Stability Issues The current state of Voice Pal is unfortunately marred by several persistent bugs that significantly disrupt the user experience. These issues range from minor playback interruptions to severe system crashes that result in permanent data loss. Interrupted Speech and Connection Stability One of the most frequent issues occurs during active conversation: the AI will abruptly stop speaking mid-sentence and fail to resume. This creates a jarring "dead air" effect that requires the user to manually exit the interface and restart the session to restore functionality. While it is difficult to pinpoint whether this is a server-side AI processing error or a client-side internet connectivity hiccup, the result is the same: a fragmented and frustrating interaction that breaks the immersion of a natural conversation. Memory Retention and System Crashes More critical is the "amnesia" bug, where the AI suddenly loses the context of the ongoing discussion. In the worst-case scenarios, the application suffers a complete system crash. When a standard glitch occurs, I can usually mitigate the damage by manually saving and exiting, then prompting the AI to recall previous talking points upon reentry. However, when the app crashes, it fails to perform an emergency save, meaning the entire history of that session is wiped clean. Having to restart a complex conversation from scratch is a major inconvenience that undermines the utility of the tool. Proposed Solution: Robust Auto-Saving To address these stability concerns, a more resilient data management system is needed. Implementing a real-time auto-save feature would ensure that every exchange is cached immediately. If a crash occurs, the user could pick up exactly where they left off without losing progress. To maintain flexibility, the system could include a simple "Clear History" or "Start Fresh" command, allowing users to choose when they want the AI to ignore previous context rather than being forced to restart due to technical failure. And lastly, the cost. This one is more of a personal gripe than a definitive problem, which is why I’m giving it three stars instead of two. I can't 100% rationalize the cost, at least for people who are not creatives. While I find the app useful enough that the $10 a month has slowly stopped bothering me, part of me would prefer a lifetime purchase option or a bit more affordability. Even though the app is extremely useful, it is still relatively niche for me because I only use it for two things. In contrast, I use Claude for almost everything, and that is free. I feel like I either need more use case scenarios or I need to learn more about the app to justify that $10 monthly purchase. If the price went down to $5 a month at some point, I think that would make it completely worth it. At $10 a month, I feel like I need more services, more updates, and fewer bugs. Alternatively, the annual subscription could be more affordable. It would be great if it was $60 a year, which I could justify much more easily than $90 to $100. It would also be very appreciated if you could include a lifetime option for $200 to $300. Again, this is just a personal gripe. I’m looking forward to the app earning that $10 monthly price tag.

I turn on and off the subscription. I turn it on because even being a well connected tech person this experience is missing and needed. But then I turn it off because it can be so much more well done. Invest in real user studies asking them what they are missing or what and add those features. For example the questions it asks can go a direction and there is no easy way to get out of it. I wish I could prompt the questions to go a different direction or refresh questions.

Where’s the VoicePal going to talk to us like Ali showed in a recent video?