Article
Custom Indicators on Your IBKR Chart, Described in Plain English
Beyond RSI and MACD: describe an indicator in plain language and have your AI build it as a sandboxed formula that renders on your live IBKR chart — no Pine Script, no code.
Built-in indicators cover the basics, but every trader eventually wants something the menu does not have. Traditionally that meant writing Pine Script for TradingView or code for a platform. On MyLinedChart, running against your own IBKR session, you can instead describe the indicator in plain language and have your AI build it — as a sandboxed formula that renders directly on your live chart. No Pine Script, no code execution, and it stays on your machine.
Beyond the Built-Ins
MyLinedChart ships the standard set — RSI, MACD, moving averages, EMA, Bollinger Bands, and volume. Those cover most routines, but the moment you want 'the gap between the 20 and 50-day average' or 'RSI but only during the first hour,' you are off the menu. On most platforms that is where Pine Script or custom code begins.
Here it does not. Your AI, connected through MCP, can invent the indicator from your description and render it on your IBKR chart. You stay in plain language; the AI writes the formula.
- Built-ins: RSI, MACD, MA, EMA, BOLL, VOL.
- Anything else: describe it, the AI builds it.
- No Pine Script and no TradingView account required.
How the Sandbox Keeps It Safe
A custom indicator is a restricted math expression over your chart's own price data — open, high, low, close, and volume — using window functions like sma, ema, rsi, stdev, highest, lowest, sum, and ref, plus basic arithmetic. That is the entire surface. There is no code execution, no network access, and no way for a formula to touch anything outside the math sandbox.
That means the worst case for a bad formula is a wrong line on your chart — not a security or account risk. And like the built-ins, a custom indicator is a capability inside the single propose_chart_changes tool; MyLinedChart still exposes nine MCP tools total, and account and orders remain read-only.
- Inputs: OHLCV of your chart only.
- Functions: sma, ema, rsi, stdev, highest, lowest, sum, ref, and arithmetic.
- No code execution, no network, no account access — worst case is a wrong line.
Examples You Can Just Say
The point is to think in trading terms, not syntax. You describe the idea; the AI proposes the formula and you confirm it onto the chart.
| What you say | What the AI builds |
|---|---|
| The 20-day average minus the 50-day average | sma(close,20) - sma(close,50) |
| A standard 14-period RSI | rsi(close,14) |
| How far price is above its 200-day average, in percent | (close - sma(close,200)) / sma(close,200) * 100 |
| 20-day rolling high of the highs | highest(high,20) |
Local, Yours, and Reusable
Custom indicators live on your device — they render on your chart and are visible to your AI through MCP, but they are never uploaded to cloud sync. They also travel with your chart's structured export, so an indicator you invented is part of the record you review later.
For the broader take on describing indicators without Pine Script, see AI Custom Indicators Without Pine Script (or TradingView). For how the whole write channel stays safe, see Let Your AI Operate Your Chart, Safely — What Changed and What Didn't, and for setup, see /mcp.
FAQ
Do custom indicators need Pine Script or coding?
No. You describe the indicator in plain language and your AI builds it as a sandboxed formula. There is no Pine Script and no TradingView account involved.
How many built-in indicators are there?
Six — RSI, MACD, MA, EMA, BOLL, and VOL — plus any custom formula you or your AI describe.
Is a custom indicator a new MCP tool?
No. It is a capability of the existing propose_chart_changes tool. MyLinedChart still exposes nine MCP tools total.
Are custom indicators saved to the cloud?
No. They are local to your device. They render on your chart and are readable by your AI via MCP, but they are never cloud-synced.
Sample Structured Chart-Data Exports
Review how chart drawings, annotations, OHLC, volume, and execution context become reusable structured data.