Day 13 of 365.
Today I ran real-world tests with the Google Analytics integration. I tested a full campaign with view, click, and reply triggers, and it worked! The next step is testing the reporting page to fetch live data directly from Google Analytics.
I also started designing a new UX/UI for our AI Data Pipeline, taking inspiration from Perplexity and some references shared here in the community.
#VibeCodingChallenge #BuildInPublic #AI #SaaS
Day 10 of 365.
Today I stayed focused on the AI Email Automation project and spent time refining the harness for the new video app POC. I conducted deep research to gather all the necessary documentation, API guidelines, and compliance rules to build it right.
On the AI Email Automation side, I spent most of the day adjusting the trigger workflows to ensure full compliance with Google Workspace Marketplace rules and GDPR guidelines.
Thatβs a wrap for today! Let's grow!
#VibeCodingChallenge #BuildInPublic #AI #SaaS
Day 9 of 365.
Today I had a meeting with my two partners to keep pushing the AI Data Pipeline project forward. I presented all product features along with the Vibe Coding harness I used to build so fast, leveraging context engineering, agents, skills, hooks, and a structured knowledge base.
Meanwhile, I worked on the new video app POC opportunity. I was able to build exactly what I wanted! I can't share the details just yet, but it is an incredible feature for a trending video production method.
Thatβs a wrap for today! Let's grow!
#VibeCodingChallenge #BuildInPublic #AI #SaaS
Day 8 of 365.
Today I stayed 100% focused on the AI Email Automation project, refining the UX and UI. I tested all features for both single emails and automated sequences with triggers, including delay, view, click, and reply actions. The Google Analytics integration worked smoothly!
As soon as I finish testing, I will shift my focus to the Landing Page to prepare for launch. Let's grow!
#VibeCodingChallenge
Day 7 of 365.
Today I focused heavily on the AI Email Automation project, polishing specific UI components. In my assessment, the design is currently at an 8/10, though I still faced some organizational hurdles along the way.
Meanwhile, I worked on the Landing Page for the AI Language Study extension to prepare for promotion. I also updated the harness with a new AI model for the Conversation module. It is incredible... by far the best model yet, enabling 100% natural conversations. The AI can even cough! Haha!
I built a Landing Page for AI Realtime Translation as well. I used the core product code to generate the actual content and mockups, leveraging specialized agents to handle the copywriting and UX writing. I also updated the primary agent mode.
Lastly, I am wrapping up the new project POC for a video app opportunity. Big news coming soon!
Two screenshots are of the AI Email Automation.
#VibeCodingChallenge
Day 6 of 365.
Today I presented the AI Data Pipeline to our team, and we are super excited to build all of its features. This app has massive potential, and we are aiming to launch it later this year. I walked the team through the core problem, our proposed solution, and the full product roadmap.
For AI Realtime Conversation, I implemented the Grok Voice API as an alternative provider to get much more realistic voices, as OpenAI sounds a bit too robotic for our use case. Next up: bug fixes and refinement.
On AI Email Automation, I am working on the core email workflow to trigger custom automations. I am still tweaking the UX/UI and will keep pushing on this.
Meanwhile, on the Umbrella infrastructure, I am continuing to build out the subscription module.
Let's grow and have a great week, everyone!
#VibeCodingChallenge
Day 3 of 365.
Today I had an amazing meeting with two new partners to scale the AI Data Pipeline. We are forming a team to target the franchise and non-profit sectors. One partner will drive business structuring, marketing, and sales, while the other takes on technical leadership and early operational support. We plan to launch in 2 months with initial customers lined up.
Big breakthrough on AI Realtime Conversation: I built our custom audio routing from scratch without third-party dependencies, and successfully ran the first live translation test. The demo turned out incredible. Still plenty to do, like stripping test features and stress-testing for edge cases, but it proves the concept for remote devs working globally, customer care, and multilingual sales calls.
For the AI Language Study extension, I made a few UX/UI fixes today, building on yesterday's massive load-time performance gains. Meanwhile, the Umbrella checkout infrastructure is almost ready, with just 2 tasks left before testing.
On AI Email Automation, the harness worked on UX/UI design. It's sitting at a 7/10 right now, so I will refine it on Monday. Lastly, I am prototyping a new POC around a fresh opportunity with new AI models. If it works, it could launch with a dedicated team in about a week.
Let's grow and have a great weekend, everyone!
Here is a look at my agent harness for Vibe Coding. I am heavily using context engineering so the AI can make better decisions on its own.
My setup features over 70 structured context files that guide the project to move faster and commit significantly fewer errors.
Every Vibe Coder needs a proper status bar in their CLI. Seeing your configurations and limits in real time is essential to monitor how the AI is performing.
In mine, I can track how many background agents are working, the active git branch, total session time, lines added and removed, warning indicators when context crosses 200k+ tokens, and even the estimated costs if I were paying per API call.
You can easily customize yours using /statusline to fit your exact workflow.
@dxiaolong It was a bloated data reading path. We were loading and checking thousands of stored words every page load just to use hundreds. We fixed how the data is stored and filtered, which wiped out the delay completely.
Day 2 of 365.
Focus: Umbrella Branding subscription strategy, strategic partnership meetings, and deep parallel progress.
Even with 4 hours in meetings, the AI kept building in the background.
What got shipped/optimized today:
AI Language Study: Cut extension load times by 90%, working on Conversation Mode, and calibrating its agent harness.
AI Email Automation: 100% focused on UI/UX redesign and DB conventions.
AI Realtime Translation: Setting up virtual audio channels for real-time PT β EN voice translation. Complex setup across Windows, macOS, and Linux, but speed is critical.
Umbrella Infrastructure: Modeled product-level Stripe subscriptions. (Stripe Brasil makes multi-currency tricky, but found a workaround).
AI Video Studio: Long-term strategic project (100% AI-based video editor for branding/awareness). Integrated MiniMax H3 for testing (first/last frame, image/video/audio refs). Deep research shows hosting open-source models can slash API costs by 10x.
AI Data Pipeline: Resurrected this previously on-hold ETL/ELT project (models can finally handle it!). Opportunity to roll out a POC this month to our first client, a Food Service franchise chain, via a strategic partnership.
Tech stack/AI today: Claude Code Max, SuperGrok Heavy, ChatGPT Plus.
Total focus remains on AI Realtime Translate and the Umbrella platform.
Planning + setting AI to execute = maximum leverage.
Let's grow!
#VibeCodingChallenge
Day 1 of 365.
Starting the journey from $0 to $1M ARR today.
Iβve prototyped 15+ AI concepts across video, audio, language learning, automations, and developer tools.
The goal isn't just to build. It's to turn the best ideas into scalable SaaS products.
Focus: Umbrella Branding with SaaS Subscription.
Parallel builds today: AI Email Automation, AI Language Study, and AI Realtime Translate.
Shipping in public. Follow along as I validate and scale each one.
#VibeCodingChallenge
Today, with AI, the biggest challenge isn't building a product. It's building a system that lets you launch the next one faster, without starting from scratch every time.
An umbrella structure provides the foundation: reusable software architecture built for scalability and internationalization, a shared audience, pre-orders, subscriptions, governance, and policies.
The real leverage isn't just in the product. It's in the infrastructure behind every launch.
I think it's AI Realtime Translation. It's almost ready, just testing, fixing bugs, and tweaking the harness. I'm planning to launch the alpha version in 2-3 weeks.
I agree that it's a lot. But with AI and a good harness, I can give them instructions and let them run for minutes or even hours in parallel.
I'm starting a new experiment:
Vibe Coding until I find Product-Market Fit.
$0 β $1M ARR Vibe Coding Challenge.
The plan:
π€ Build fast
π Launch fast
π§ͺ Test multiple ideas
π Share the numbers
β Kill what doesn't work
π Double down on what does
No perfect idea. No endless planning.
Just build β launch β learn β repeat.
Until one hits Product-Market Fit (PMF).
Then β $1M.
π 09/09/26
#VibeCodingChallenge
I built my own English teacher in two weeks.
It's a Chrome extension. While I read the web, it quietly collects every word I don't know... from my own pages, my own work, not a textbook. I choose what goes into study.
Then I just talk to it. Real-time voice, both ways.
And here's the part I'm proud of: the teacher cannot use a word it hasn't taught me. Every sentence passes a vocabulary gate before it's ever spoken. So it teaches, then asks... never the other way around. When it asks a question, it also gives me the sentences I can answer with. No quizzes. It's a conversation.
Every word I say back becomes a flashcard, and spaced repetition takes it from there.
Listen. Read. Write. Speak. Converse.
Built on my own dev framework. Which is the only reason two weeks was enough.
Another new product launching soon.
#llm #ia #app #claude #grok