Are there trading limits (per-transaction or daily) when using an AI agent?

Yes — Pluang enforces system-level spending limits on orders an AI agent places on your behalf: a maximum cap per order (transaction), per day, and per asset class. These limits are enforced server-side — locked from Pluang's backend, not something the AI agent or you can override or bypass — as confirmed in "Is Agentic Trading on Pluang safe?". The three ceilings apply at the same time, so an instruction that clears one can still be refused by another: an order within the per-transaction cap may be blocked once your daily total is reached, and unused room in one asset class does not release room in another. The agent cannot raise, waive or negotiate any of them, and it cannot report a limit as lifted when it is not — if a cap blocks an order, the order simply does not go through, and you can place it yourself in the Pluang app instead. Pluang does not publish the specific figures.

Pluang Agentic Trading's spending limits work as a server-side control, separate from — and layered on top of — whatever access-scope permissions you've granted the agent (see "How does Pluang control what an AI agent can access or do in my account?"):

• A per-order cap — a maximum amount any single transaction the agent places can be worth, regardless of what the agent was instructed to trade.
• A daily cap — a maximum cumulative value of orders the agent can place within a day.
• A per-asset-class cap — a separate ceiling for each asset class the agent can trade (e.g. US Stocks, Indonesian Stocks, Crypto), rather than one flat figure applied across all of them.

These caps are enforced entirely on Pluang's backend, not on the device or within the agent's own logic. The AI agent cannot raise, override, or work around them. Pluang does not publish the specific figure for each cap.

The reasoning behind having these limits at all is straightforward: even with a well-scoped permission grant, a single misunderstood instruction, a bug in the agent's logic, or a runaway trading loop could otherwise place trades at a damaging size or frequency. A hard, server-side ceiling bounds that risk regardless of what goes wrong upstream — a backstop that doesn't depend on the agent behaving correctly.

Related questions:

Q: Why does Pluang enforce these limits server-side instead of just relying on permission scopes?
Permission scopes control *what kind* of action an agent can take (e.g. place US Stock orders, but not withdraw funds); a spending limit controls *how much* damage any single allowed action — or a burst of them — can cause. Enforcing the cap server-side also means it can't be tampered with from the agent's side or the app itself; it holds even if something upstream misbehaves.

Q: What happens if my agent tries to place a trade above the limit?
The order is blocked or rejected server-side before it executes.

Q: Will the exact cap amounts be the same across every asset class?
No — caps are structured per asset class, so each asset class (e.g. US Stocks, Indonesian Stocks, Crypto) has its own separate ceiling rather than one flat figure applied across all of them.