# Architecture

AeroNavigator is not a simple AI wrapper. It is a sophisticated Asynchronous Multi-Agent System running on NetMind’s decentralized compute network.

To achieve high-frequency execution and eliminate the latency of monolithic models, we split the workload across specialized agents that operate in parallel.

### The multi-agent system

Workloads are split across specialized agents. Agents run in parallel to reduce latency.

#### 1. Data Collector (Scout)

The Scout gathers market data before you ask. It scrapes, cleans, and vectorizes on-chain and market feeds.

**Function**

* Filters noise and removes duplicates.
* Stores relevant context for fast retrieval.

**Why it matters**

Pre-processing reduces latency at query time.

#### Research Agent (Analyst)

The Analyst activates when you query a token or pool. It retrieves the most relevant data fragments and on-chain signals.

**Output**

* Sentiment scores.
* Qualitative impact notes.
* Fundamental and market pattern context.

#### Decision Engine (Trader)

This is the execution core. It fuses sentiment with real-time on-chain data via the Sugar API.

It also uses technical indicators such as RSI and MACD.

**Capability**

* Accesses 100+ MCP (Model Context Protocol) tools.

**Result**

* Constructs a final transaction bundle for you to sign.
* Examples: swap, LP add/remove, gauge vote.

### Four-layer stack

#### Layer 1: Decentralized compute (NetMind Power)

At the foundation lies NetMind Power, a decentralized network of GPUs providing the
\
computational horsepower for complex inference tasks. This architecture delivers a %
\
cost reduction compared to centralized providers (AWS, Google Cloud) while ensuring data
\
privacy and resilience.

#### Layer 2: Inference engine (multi-agent reasoning)

The reasoning layer uses a specialized multi-agent system with parallel processing for
\
faster response times and more accurate reasoning.

#### Layer 3: Data layer (Sugar API + MCP)

Real-time data is critical for execution quality. AeroNavigator bypasses traditional indexed
\
subgraphs (which lag by seconds to minutes) by using the Sugar API for direct, non-indexed
\
access to Aerodrome's liquidity state.

**Sugar API**

* Packs pool data into one call.
* Includes TVL, APR, emissions, user positions, and gauge status.
* Enables real-time APR calculations.

**MCP tools**

* Standardized access to on-chain data.
* Access to external APIs and communication channels.

#### Layer 4: Application layer (interface)

The user interface is deployed as a Base Mini App and on Farcaster, combining natural
\
language chat with visual components (Range Cards for V positioning, allocation
\
visualizations, etc.)

### Why Base

Base is the execution layer for AeroNavigator. It provides:

* Low fees for automation and frequent rebalancing.
* Aerodrome’s ve(3,3) liquidity model.
* Strong ecosystem incentives and builder support.
* Farcaster as a native identity layer.
* Mini Apps for distribution.
* Coinbase custodial and non-custodial rails.

AeroNavigator is Base-first. It is designed around these primitives.

### Next steps

* See what’s already shipped in [Season 0 features](https://aeronavigator.gitbook.io/aeronavigator-by-netmind/product/season-0-features).
* Read the constraints in [Security](https://aeronavigator.gitbook.io/aeronavigator-by-netmind/trust/security).


---

# 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://aeronavigator.gitbook.io/aeronavigator-by-netmind/product/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.