Roman of Nodarama. Introducing a category defining AI Code Recorder, from vibe to viable.
Thanks Chioma for the warm introduction to Saas Hive, glad to introduce Nodarama Verbatim here.
My Name is Roman, and I have been 'vibe coding' since before it had a name, and in 2025 I was in the top 0.1% of ChatGPT users, which I was informed in November in their 'unwrapped' surprise and delight moment. I don't think I was delighted, but nobody can say I have insufficient experience. :)
Nodarama Verbatim is an AI Code Recorder, which means, it facilitates easy transfer of CODE from your AI Chat, to your local project VIA a Browser Extension, and connected Code Recorder, as well as sharing your base code, from your system, with AI chat. AKA all you need to do is plan and prompt, AI codes, and Verbatim captures and follows those instructions.
By automating this process you reduce the friction associated with, trouble shooting formatting errors, and lengthy waits and token costs associated with a full file rewrite, for every small change or revision to your project. Nodarama Verbatim, helps vibe coders, create original software, leveraging AI, but preserving authorship.
If you’re into vibe coding, I’d love for you to try it. If you message me and are open to giving private feedback about how it works for you, I’ll extend 1–3 free months of unlimited use.
Thank you!
Xplania
Your AI-powered travel companion, from planning to memories ✈️
Comments (9)
Welcome to the community, Roman. May I ask what is the advantage of using Nodarama Verbatim over the way of me vibe-coding inside the IDE? 🤔
HI Sergey, thanks for the welcome & the question. The Verbatim coder itself won't necessarily offer much value to someone proficient in coding already using full IDE's, it's better suited to vibe coders who have learned the way they do it today is not strong enough, looking to start advancing their involvement. That said, lots of developers are vibe coding, and I would say its most similar to Codex and Claude Code, and the main benefit outside of those mentioned already would be the ability to use with any model you can access through Chrome including locally run models. Essentially you get 1 click import from your CHAT to your Project.
Possibly some of the smaller features could be compelling though but they aren't breakthrough features or anything, Project packages as well as recording contracts to bring a new chat up to speed quickly, Auto collapse view in the ultra light editor. Nodarama is a full ecosystem, there's a node launcher for small quick projects someone makes with verbatim, organizing and consolidating in a simple UI. So building nodes for functions. Which is the same node contract that is used in Atelier which is the kernel powered automation engine. Verbatim is one small part of the system extracted to release first. Stay tuned :)
Thanks for introducing Verbatim, Roman. What kind of project do you think people should try first to really appreciate the difference Verbatim makes?
Great question, there is a suggested onboarding process, recommending a personalized launch pad, you get to make it however you want, it doesn't need any unique or challenging features, but you will practice some foundational planning skills right off the get go. That you will need if you continue to build.
"Preserving authorship" while still leveraging AI is an interesting angle, it's basically a trust problem: proving a human was actually driving the plan, not just accepting whatever the model output. I ran into a version of this building FounderFlow. Once an AI is producing a lot of your day-to-day output (in my case, insights across three businesses), you need some way to know which parts are solid and which are the model filling gaps, otherwise you stop trusting your own tool. Does Verbatim keep any record of the original prompt/plan alongside the captured code, so you can trace a specific change back to the instruction that produced it?
Yes right now there are a few of the earliest methods I used to do that, primarily project note lists, and a self populating filemap/ second line scope definition on every file, and a quick easy way to edit them all at once.
There are logs for all the automatic AI calls for tracking but the full conversation is not captured when you have an external conversation with AI. But you pose an interesting consideration, perhaps users will want to capture their chat history as they go. I know one pain point is the lack of time stamps. If it was captured they could get added easily.
The logs are thorough though, every change would have a comment around the objective. But currently no, not a full 2 way dialogue capture, at this point in time. (That could change really fast as I think it through further)
Roman, that timestamp gap you mentioned is a good catch, that alone would let someone reconstruct intent after the fact even without a full transcript. The objective-per-change logging sounds like a reasonable middle ground for now. Following this one, curious to see where the rest of the ecosystem goes.
I'd love to tell you now, because I am a blabber mouth, but AI produces things so fast, and I am not the fastest (commmon capital, I need a team HAHA) so it would be suicide to share. LOL :)
Lol Roman, that's the most honest answer I've seen on here. Keep building!
Sign in to comment or upvote.