TradingView AI Copilot: Current Features and How to Check Its Output
TradingView AI Copilot now runs inside Supercharts. Older descriptions of a Chrome-only assistant with a 15-request daily allowance no longer describe the current product. The useful question is how to check an assistant's work before relying on it.
How TradingView AI Copilot Works
TradingView's native Copilot announcement describes a separate add-on for paid plans on web and desktop. It uses chart context and TradingView market data, can change chart settings, manage alerts and watchlists, write and apply Pine Script, and explain strategy backtest metrics. The announcement specifies monthly usage limits and no mobile-app availability. Check the checkout screen for the current price and allowance.
The earlier extension announcement now says that the browser extension is no longer actively supported and will eventually be discontinued. Start with the native product rather than installing an extension from an old tutorial.
Where the Copilot Needs Verification
An answer can sound precise while referring to the wrong exchange, session or timeframe. Before prompting, record the full symbol, chart interval, session setting and date range. After a chart action, compare those settings with your request.
For a reproducible check, ask the assistant to add a moving average with a specified length. Open the indicator's settings yourself and inspect its inputs. Repeat with a different length. This checks whether the requested action occurred; it does not measure forecasting accuracy.
Checking Generated Pine Script
Code generation is available, so choosing another tool should depend on your workflow rather than a claim that Copilot cannot write Pine Script.
Before generating a strategy, write down entry conditions, exit conditions, order timing, position sizing and trading costs. Then inspect whether the output implements each condition. A script that compiles can still calculate a signal differently from the intended rule.
Keep the prompt, source version, compile result and unresolved issues together. If you compare Copilot with Pineify or another assistant, give each the same specification and judge the resulting source against the same checklist. A useful comparison reports the sample and failures; a single successful script cannot establish a general accuracy rate.
Checking the Data Behind an Answer
Ask for the reporting period and source behind each material figure. A quarterly revenue number, a trailing annual figure and an analyst forecast answer different questions. Open the issuer's filing or earnings release to resolve discrepancies before building a trading rule around them.
For price-based claims, distinguish a historical close from the latest trade. Record the exchange and timestamp alongside the value. A response without those details is incomplete evidence, even when the number resembles the chart.
A Practical Evaluation Checklist
| Task | Evidence to inspect | Reason to repeat the check |
|---|---|---|
| Change a chart | Symbol, interval and visible settings | The chart context changes |
| Explain a financial figure | Filing, period, units and source link | A new report is published |
| Generate a script | Source, compiler result and rule checklist | Code or requirements change |
| Explain a backtest | Trade list, settings and calculation period | Costs or history change |
| Create an alert | Symbol, condition, frequency and expiry | The underlying setup changes |
This is a suggested evaluation procedure, not a report of a completed benchmark.
From Research to a Tested Strategy
- Define the question. Separate a descriptive question about the chart from a rule that could generate trades.
- Write the rule. Specify what information must be available when a signal occurs and when an order should execute.
- Inspect the generated source. Resolve compiler errors and compare the implementation with your specification.
- Review individual trades. Check several entries and exits on the chart instead of judging only the headline return.
- Vary the assumptions. Examine whether costs, a different period or a nearby parameter value materially change the result.
- Keep an evaluation record. Save the source and settings so a later result can be reproduced.
For more detail on test design, see our guide to validating strategies before live trading.
Example Tasks to Try
For chart setup, request one visual change and verify it before asking for several changes together. For research, ask for one reported metric and trace it to its source. For coding, start with a rule you can calculate manually on a few bars.
These small checks reveal specific problems: a mistaken timeframe, mismatched reporting period or ambiguous entry rule. They are more useful than asking whether an assistant is generally “good at trading.”
Choosing a Tool for Your Workflow
Use the same acceptance criteria for every assistant. Consider where you want to edit source, how you retain versions, whether data sources are visible and how much manual correction is necessary. Confirm current plan terms before paying for features you expect to use.
Pineify is an option for readers who want a separate script-building workflow. Compare it on your own specification; this article does not establish that either product is more accurate or more profitable.
How to Get Started
Open a chart while signed into an eligible account, select the Copilot icon in the right panel and review the access offer. After enabling it, run one of the checks above on a saved test layout. Keep changes small enough to inspect and undo.
Frequently Asked Questions
Can TradingView AI Copilot write Pine Script?
Yes. The native product supports Pine Script generation and application. Inspect the resulting source and test its behavior against your intended rules.
Does the old 15-request daily limit still apply?
Do not use the old extension allowance to evaluate the native product. Review the current monthly allowance shown in the access offer.
Does a successful compilation prove a strategy works?
No. Compilation does not establish that the strategy follows your specification or that a historical result will persist. Review the code, trades and assumptions separately.
How should I compare Copilot with Pineify?
Give both tools the same written task. Save the output, count the corrections needed and verify the final behavior. Report those observations without generalizing from a small sample.