TRUE AI: AI trading chat that combines research and execution

True Labs built TRUE AI on Solana as a non-custodial platform that links AI market research, on-chain trade execution and optional automation in a single chat interface. The company raised a $10 million seed round in 2025. The app is designed so users keep control of their wallets while using the platform’s tools.

The research layer pulls live prices, order flow, liquidation heatmaps, technical indicators, charts and market news into one view. The assistant reads those inputs and returns plain-language explanations in multiple languages. Users can upload charts or documents to the assistant for analysis or summary.

The trading layer supports non-custodial token swaps routed through Jupiter, perpetual futures on an on-chain DEX, copy trading and cross-chain bridging. Card payments are available through a third-party provider; identity checks apply only to fiat on-ramps handled by that provider.

A typical chat interaction begins with a command such as “Long BTC.” The assistant parses intent — direction, asset and leverage — then gathers context across timeframes, order flow and indicators. The platform highlights risk metrics, including likely liquidation zones and suggested stop-loss levels, and shows how a new position would affect existing holdings.

Execution is user-driven. Every leveraged trade, limit order and stop-loss tied to the DEX must be placed and signed from the user’s wallet. The assistant may execute a simple token swap only when a user explicitly requests it in chat. Automation runs inside permissions set by the user and can be revoked at any time.

Autopilot agents are optional and permissioned. Users set approved assets, spending caps, stop-loss rules and an expiry date for each agent. All automated decisions are logged and explained so users can review why and how an agent acted.

TRUE AI describes itself as “agentic,” meaning the system chains multiple reasoning steps and uses external tools and data sources. For developers and advanced users, a read-only Model Context Protocol (MCP) connector allows compatible AI assistants to pull authorized market and account data without any ability to move funds.

The True Labs team has published research on additional safety layers for autonomous agents, an approach it calls SAE. Platform documentation states limits: the AI can be incorrect, automation does not remove market risk, and self-custody places responsibility for signed transactions with the user. Features and token mechanics may change as the platform develops.

The product targets active on-chain traders who prefer to retain custody of assets while using integrated market intelligence and permissioned automation. The company does not present the platform as a hands-off investment service or a guarantee of returns; users are advised to perform their own research and manage risk accordingly.