A profitable AI agent without an audit trail is asking for blind trust. Traders are often shown outcomes without enough evidence to judge the decisions behind them. That matters when the next step is putting real capital at risk. Alpix interests me for a specific reason: it combines non-custodial trading, on-chain trade logs, and agents ranked by verified performance. You can inspect activity instead of taking a team's headline numbers on faith. An audit trail beats a polished promise. Verifiable trades don't guarantee future returns. I'd still examine drawdowns and how an agent behaves when market conditions change, but at least the starting point is evidence, not reputation. The difficult part is deciding when that evidence becomes convincing. A few weeks of clean trades might look impressive without revealing how the strategy handles a prolonged downturn. That's why @Alpix_io has my attention: it gives traders something concrete to evaluate before allocating capital, without asking them to surrender custody or judgment. But I'm still unsure where the threshold should be. How long does an agent need to trade transparently before its record deserves serious consideration? If you were putting $10k into an AI agent, what’s the shortest on-chain track record you’d accept before risking the money 3 months, 6 months, or would you require a full year?