TrendSpider Sidekick vs TradingView AI Copilot: what the published tests and price lists show

Chart copilots promise the same thing: describe what you want in plain English and the AI draws the levels, writes the indicator and sets the alert. The two most visible ones now sit inside the two charting platforms many US traders already pay for. Here is what published tests and the vendors' own pages say, as of October 14, 2026. Plan details for both platforms are kept current in our TradingView and TrendSpider tool pages.

What each one is

TradingView AI Copilot started as a free Chrome extension in public beta on April 2, 2026, then became a paid add-on built into the chart on September 28, 2026, according to TradingView announcements cited in the review by Kagels Trading. It reads the open chart, draws levels, writes Pine Script, looks up prices and creates alerts. It runs in the browser and desktop app, not the mobile apps, and requires a paid plan (Essential, Plus, Premium or Ultimate).

TrendSpider Sidekick is an assistant inside the TrendSpider platform. Its pricing page lists it behind 'AI Strategy Analysis' and 'Machine Vision Analysis' features and names the models it can call: Sonnet 4.6, Gemini 3.1 and GPT 5.4 on every tier, plus GPT 5.5 and Opus 5.0 on the two top tiers. TrendSpider's blog says Sidekick can build a custom indicator step by step and check the rendered chart after each step.

The price lists, side by side

Sources: TradingView pricing, TrendSpider pricing. TrendSpider's low figures are introductory: its terms say the discount covers the first invoice only. TradingView's page also lists 'AI screener requests' of 100, 150, 250 and 500 by plan, without stating the period. Kagels Trading reports the add-on's base value in TradingView's page code suggests $7.50 in the US, but did not verify it at a US checkout, and neither have we.

The Copilot test: seven tasks, mixed results

Karsten Kagels bought the Copilot with his own money and gave it seven checkable tasks on September 30, 2026, recalculating every number.

Two details stand out. The indicator that never fired came with a green check mark: 'compiled' meant no syntax error, not a working rule. And the Copilot quoted a rate-hike probability of about 70% without a timestamp, while Yahoo Finance reported roughly 35% that morning citing CME FedWatch nContext: n=1 query · Sep 30, 2026 · Kagels Trading review, citing Yahoo Finance and CME FedWatch · margin: not disclosed. Usage is shown only as a percentage; it went from 92% to 89% over the seven tasks nContext: n=7 tasks · Sep 30, 2026 · Kagels Trading review · margin: not applicable, and TradingView does not say what 100% equals.

The Sidekick reviews: useful, metered

No published Sidekick review we found ran a scored task list like the Copilot test, so the evidence is softer. StockBrokers.com ran a full workflow from Sidekick idea to backtest to bot and reports using 18 messages in a couple of hours, against 25 included per month. The reviewer says Sidekick flagged when a backtested strategy's return came from a few outsized wins, and that most strategies built with its help were still flagged as not ready for launch.

FullStack Alpha's comparison adds a metering detail: a Sidekick Deep Research request spawns one sub-conversation per symbol, so a five-ticker request uses six messages. Neither vendor publishes an audited accuracy rate for its assistant.

How to read this

• Data access is the real feature. Both assistants work on their platform's own data rather than the open web. That helps with prices; it did not prevent a stale macro figure in the Copilot test.
• Generated code needs a known-answer test. Run any AI-written indicator on a period where you already know where the signals should be.
• Count the meter before the subscription. 25 messages can go in an afternoon, and a percentage gauge gives no count at all.