Building a trading indicator without writing code

A friend tossed out an idea on the way to a ski trip. I opened Claude on my phone. By the time we crossed the border, a working TradingView indicator was running. No Pine Script experience needed — just a clear idea and the right tool.

Daniel Plomp

03 Apr 2026 — 5 min read

While driving to the Alps, a friend casually mentioned the trading levels he checks each morning. Previous day's high and low, the open, the close — the four reference points that most active traders check before doing anything else.

"There should be a clean indicator for that," he said. "Most of what's out there is either bloated or ugly."

I had my phone in hand. By the time we crossed the Austrian border, the first version was almost running on TradingView.

The idea was simple

Plot the previous session's Open, High, Low, and Close as horizontal lines on a chart. That's it. Traders call these OHLC levels, and they serve as natural support and resistance zones. The previous day's high is where sellers stepped in. The low is where buyers showed up. The close tells you the market's final verdict.

Most charting platforms offer this feature, but they often fall into two extremes: either too basic with no customization options or too complex with endless settings and nested menus. The gap in the middle — a clean, focused, well-designed indicator — was wide open.

Building it without writing code

I don't write Pine Script. I'm a web developer. I build Laravel applications and CMS platforms. TradingView's scripting language was entirely new to me.

But I can clearly describe what I need, and that was enough.

I opened Claude on my phone and described the indicator I had in mind. Not in technical terms — just in plain language. "I want an indicator that shows yesterday's open, high, low, and close as horizontal lines. The user should be able to pick daily, weekly, or monthly. Each line should have its own color and style."

Within minutes I had working Pine Script code. Not a rough draft — actual code I could paste into TradingView's editor and run immediately.

The iteration loop

The first version worked, but it wasn't good enough. The labels overlapped when price levels were close together. The lines started from the current bar instead of from where the previous period actually began. Small things, but the kind of details that separate a polished tool from a quick hack.

So I kept describing what I wanted to improve. "When two labels are close together, they should spread apart and show a dotted line back to the actual price level." That's not a trivial feature — it requires sorting all active levels, detecting collisions, and adjusting positions dynamically. It was the kind of algorithm I could understand conceptually, but implementing it in an unfamiliar language would have taken hours.

AI handled the implementation. I handled the product thinking.

This became my routine: describe the behavior, test it, identify improvements, and iterate. Each cycle took a few minutes. After a handful of iterations, the indicator was doing things I hadn't even planned for initially — like non-repainting data handling (a critical concern for any trading tool) and configurable zone shading between the high and low.

Here's what this means

This story highlights how AI can assist in the development process by bridging the gap between idea and execution. It's a story about AI as a force multiplier for people who have ideas but lack expertise in a specific domain.

I understood the problem space. I knew what traders need because I've spent time looking at charts. I could evaluate whether the output was correct, whether the UX was right, whether the feature set made sense. What I couldn't do was write Pine Script from scratch.

That gap between "knowing what to build" and "knowing how to build it" used to be insurmountable without either learning the language yourself or hiring someone who knows it. AI significantly narrows that gap. While you still need to think clearly about your goals and critically evaluate the output, it allows you to transform an idea into a published product in just one afternoon.

The result

Previous OHLC Levels is now a free, open-source TradingView indicator. It supports daily, weekly, and monthly periods. Every level has independent color, style, and visibility controls.

Labels are smart enough to avoid overlapping. The code is clean, well-commented, and published on GitHub for anyone to learn from or extend.

It started as a conversation in a car. It became a real product because the barrier between idea and execution has fundamentally changed.

The bigger picture

If you have domain knowledge in any field — trading, medicine, logistics, education — and you've ever had an idea for a tool but dismissed it because "I'm not a programmer," that barrier is lower than you think.

Describe what you want clearly, iterate on the output, and trust your expertise in the problem space. The code is just the implementation detail.

I call this approach vibe-coding. You don’t have to know the syntax. You need to know the problem. AI handles the translation from intent to implementation. You handle everything else that actually matters: the idea, the user experience, the domain expertise, and the judgment calls.

The future belongs to people who can think clearly about problems and communicate solutions effectively. The syntax is becoming optional.