I thought when I set out to create my first app that it would take a few months. If I had just stuck with Cursor, this might have been the case. For some reason, I felt the need to do things differently. Oh, I began by using all of the apps, testing each of them for a few days to see what each created, essentially starting from scratch with each. I wanted to see what each of them created and to find out which worked best for me. I tried VO, Bolt, Lovable, Replit, Cursor, and Windsurf. They each did things a little differently from the others.
I settled on Replit, not because what it was designing was any better than the others, but because I liked their interface, and I had a discount. However, I quickly learned that the AI running it was not the greatest and would create boring styles. Despite that, by the end of March, I abandoned the others and focused on Replit. Shortly thereafter, I created the company.
It was around this time that I realized that while I liked the style of what I was creating, I did not love it. It felt like it wasn't what I really wanted. So, I began creating images for the pages of the app that I actually liked. This is why some parts of the app will not have a hover effect or change colors when you click a button, because some of the pages are background images with invisible buttons laid over the background images. This made the process more tedious as I had to spend time getting those invisible links in the right place. This also meant I was doing work away from Replit to create the images, something that would become a pattern. What do I mean?
By the end of April, I had exceeded my allotted usage for the month because the agent would often send me in a loop where it may or may not fix one thing, but it seemed to always break four other things that I then had to fix. This worked out well for me, not because I like wasting my time and money, but because I had a ChatGPT account and o3 had come out in December, Deep Research came out in February, and o4-mini-high released mid-April. So, it made sense to run my code through those two models and Deep Research until my Replit account reset, so I did not get stuck paying more to Replit. In hindsight, I probably should have shifted to Cursor or Windsurf at the time, but I was liking the results I was getting from taking my code from Replit and running it through the OpenAI models before pasting the results back into Replit to test them.
I quickly grew to like it better than just trusting Replit's overly expensive and flawed agent. This way, I actually felt more in control, and most importantly, I was learning about the app-building process. Not that I would ever attempt to build an App without AI, but I began to understand how the app actually worked. This agreed with me in a deeper way because I have always enjoyed deconstructing things (toys, household items, anything mechanical) and then putting them back together. So, by May, I was essentially just using Replit as a code viewer to test what I had done over on ChatGPT. Believe me, o3 and o4-mini-high also made tons of mistakes and took me down rabbit holes that wasted loads of time, but I trusted the outputs more than Replit, whose agent would break more than it would fix. I still used the Replit agent on occasion, but I stopped relying upon it as my daily driver for building the app. So, that became my process, and weeks turned to months, and my hopes of releasing the app by the end of May faded.
A lot has been learned since April. It was my choice to go this route, and there is no way I would have as much understanding of the process as I have now had I either just let Replit or Cursor make the app for me without my hands-on involvement. That said, I won't put myself through all of that again with my next app. I'll be able to use what I have learned to let Cursor or one of the countless other vibe coding tools do most of the driving on the next app. I no longer want or need as much control over every step in the process. And I can't afford to take the time to do it again the way that I have. I am grateful for what I have learned, but I need things to go much faster from here on out.
There were personal issues that arose during the past 6 months that cost me time, so I can't blame my languid building pace just on my own process. I won't detail those here, as I prefer this space to discuss my work and not my life, but these personal issues affected me at a deep level that I have never had to deal with before. How much time was lost in my attempts to deal with them in one way or another? That's tough to say, as my peace of mind was rocked by one, and my heart was broken by the other. I was definitely thrown off my game by both. Could I chalk up a month or more of delays to them? Maybe, but I kept working through it all, so it is hard to tell. When I do an autobiography, I can talk more about those issues.
Anyway, the Beta version of the app was finally ready in late September. This is the Replit-hosted version. Back in July, I began trying to get the app off of Replit and focused on moving it to Vercel. Then, after a few weeks of work, I realized that I could do it on Cloudflare, which meant I then had to do things differently. Fortunately, GPT-5 Thinking was released and has been much better than o3 and o4, who were good but not great. Even GPT-5 Thinking has had issues, but not as many as its predecessors. I have also incorporated CODEX into my workflow this month, and it is even better at times than GPT-5 Thinking and integrates into Github. Yes, thank you.
My goal for the Web App Paid release was initially May. I can't believe it has taken so much longer. The initial funding I raised to get the company started has all but dried up, as I had to keep pushing the release date. All the other plans I had made at the beginning of the year had to be put on hold after July, as I missed my second target for the release of the app. I heard people saying that you should not voice how long or how difficult it was to release something, but rather how quickly you were able to do so.
The frustrating part has been watching people whip up apps in a matter of minutes as all the tools for building apps have continued to improve, while I have been tediously fixing every little bug and not being able to focus on all the other projects I had in mind for the year.
Cry me a river, right? Well, no longer. By this time next week, the paid version of the web app for 1st Draft will be available, and every person with a story idea and a desire to tell it can use my app to their heart's content. Then I can begin to shift my focus to include other projects... finally.
The purpose of the 1st Draft App has never been to replace the amazing writing tools that are out there. It has been to provide a very simple way for anyone to tell their story that doesn't involve sitting at a keyboard. I could have cloned one of the many writing apps out there that offer people dozens of options for writing their stories, but with so many buttons that it becomes confusing. What I wanted to create was something that you can log into and just begin talking to your assistant about the stories you want to tell. We all have them, but we don't all have the time, training, or patience to sit in front of a computer to tell it properly. No longer.
1st Draft is a true voice-first writing app that allows anyone to start telling their stories from the first minute. No ads, no distractions; just you and the helpful and friendly assistant. Something that I love using because I too do not like being a slave to a keyboard and can finally be out for a walk or hiking, and simultaneously be working on developing a story while touching grass. With the 1st Draft App, you don't have to be stuck inside staring at a screen. You can be lounging around or out and about, and comfortably discuss your story ideas with the assistant, who will help you shape your story with you. It really is liberating.
I tried to make the app as simple as possible. Press Start and begin chatting with your assistant, who can be a sounding board, guide you into shaping the story, or even take what you talk about and begin writing a version of the story for you. Not sure what to say next and need time to think? Press Pause, collect your thoughts, and continue when you are ready. Don't want to listen to all that your Assistant has to say, or want to stop the conversation, just press the Stop button. You'll be sent to the Chat Workspace Page where you will see the readout of your conversation. You can browse through your Chats or switch over to the Documents Workspace Page and start typing. Or jump back into the conversation with your assistant.
For me, it changes my process. I used to scratch down ideas on a sheet of paper while out and about or out for a walk. Now, I can flesh out an idea and start building it before I get home. It speeds things up and ensures that I'll be able to read what we were talking about. This is a big deal because I have some chicken scratch for handwriting, especially while trying to walk.
If you are like me, there will be times you want a snappy back and forth with your assistant, and there will be other times that you want to try and flesh things out a bit more. The Settings Page provides options that control how the agent behaves. Whether you want to have a quick back and forth, or if you want to get deep into the details, or have your assistant create longer sections, you can do that too. There is a guide that explains it all. But if you just stick with the defaults, you can't go wrong either.
Another thing that was important to me was keeping the price low while also providing a tool that everyone can use and benefit from. You will have the choice of using the free tier, which will help you get a feel for how things work, but the Starter Plan is the way to go for most people. And it only costs $5 a month or $50 a year. Some people will never reach the monthly limits with the Starter plan, but if you are like me and you have a ton of stories you want to tell, then the Pro plan is the way to go. And if you ever reach your limits before the end of the month for either plan, you can always purchase a boost that provides extra time to work with the assistant. Even if you exceed your limits and don't want to pay more, you will continue to have access to your chats and continue to be able to work on your documents, if you don't mind typing.
1st Draft is easy to use and will keep you focused on storytelling. Whether you are young or old, have never written before or have written several books or screenplays like myself, the 1st Draft App is the place to start when you have a story you want to tell but don't want to have to sit and stare at a computer screen to tell it. Just press the Start button and speak it into existence. Just like our ancestors used to do around the campfire at night. It really is that simple and I can't wait to help you tell your stories.
Head on over to 1stDraftApp.com and get started for free, and by this time next week you'll be able to dig even deeper with even more access for a very low price. Cheers! And happy storytelling.
