# Core Architecture

**Agentic Framework:** At the core is âTARS (aarna Tokenized Autonomous Rewards Strategies), a multi-agent system with specialized Yield Curation and Execution Agents that analyze Pendle PT markets in real-time and orchestrate redemptions with minimal slippage. These agents coordinate through a graph-based orchestration layer, integrating LLM(Gemini)-driven reasoning and aarnâ’s proprietary alpha 30/7 AI signal model for adaptive decision-making. This architecture distinguishes aarnâ from static, rule-based optimization systems by enabling true autonomous yield management with explainable, auditable onchain reasoning trails.

**Vault & Token Dual Flow:** Users can either acquire aarnâ tokens (which direct 70–80% of treasury for delegated agentic strategies, with the rest in liquidity, insurance, and incentives), or deposit directly into onchain atv vaults with stablecoins to mint atv shares. 

**Vaults are fully modular:** each can host a fixed-yield strategy, a DeFi index, a stablecoin ladder, or an institutional sleeve, all managed by agent logic under protocol, market, and governance controls.

**Policy & Risk Management:** Allocation bands, diversification rules, risk triggers, rebalancing frequencies, redemptions, and insurance mechanisms are governed by smart contracts - visible, auditable, and upgradeable only via onchain community votes. Insurance, agent error cover, and risk signaling are embedded at the protocol level and visible to every stakeholder.

<figure><img src="/files/vwwzqS0Nbjk7R5lF7ioC" alt=""><figcaption></figcaption></figure>

the âtv infrastructure and the âTARS agent layer, together integrate every major DeFi venue a treasury actually needs, DEXs, lending markets, yield markets, perpetuals, structured-yield rails, LLM/AI infra, data hubs, and onchain data sources. This stack turns fragmented liquidity and scattered opportunities into a single composable system. The mobile dApp plugs directly into it, letting users access automated allocation, structured vaults, and policy-driven execution without touching the underlying complexity.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.aarna.ai/aarna-the-agentic-onchain-treasury/aarna-aot-design/core-architecture.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.