SERA has implemented embedding technology to improve the precision of data and task routing to agents. By converting user queries into explicit tasks, SERA uses numerical representations of text to direct these tasks to the appropriate agents. In a recent application, SERA utilized staking endpoints to extract real-time APY snapshots from validator APIs and staking aggregators, covering aspects like lock-up periods, commissions, and compounding behaviors. Additionally, SERA follows a "reflection to LSTs" prompt path, mapping user reflections to specific Solana Liquid Staking Tokens (LSTs) such as mSOL and stSOL, and explaining the actual flow of yields. The system also queries airdrop and aggregation information, providing Solana airdrop guides, wallet documentation, and ecosystem data streams relevant to SOL holders by the end of 2025. This approach allows SERA to break down technical backgrounds into isolated segments, enabling its reasoning LLM to understand each part from different perspectives and synthesize a more comprehensive answer.