AdEx AURA is a personal AI agent framework and recommendation engine that leverages and analyzes publicly available data from Ethereum and Layer 2 blockchains. It processes the data through a large language model (LLM) to generate personalized recommendations for actions or applications based on user behavior. In simple terms, AdEx AURA looks at the users’ blockchain activity, e.g., where they’ve sent funds or what apps they’ve used, and suggests actions or services that might interest them based on that behavior.

For initial context see this: https://www.adex.network/blog/introducing-adex-aura/

AURA is a unique product that blends together DeFAI and advertising, by being a recommendations engine for earning opportunities. The main design goal of AURA is to simplify Web3 investing and financial strategies, thereby making DeFi truly accessible.

AURA is very different from other recommendation and native advertising solutions in two distinct ways:

• Incentives flow to the user, rather than the publisher. Everyone is implicitly an advertiser, as long as their app has value to the user.
• AURA leverages AI to introduce human-like logic into the recommendation process, rather than simply using an old-school targeting algorithm. AURA will produce an answer of the question "which opportunities could be most valuable for this user" and consider all the available data from different angles

AURA is also different from AI-based DeFi automation solutions, in that most of them take a human-readable prompt and output instructions. AURA doesn’t need a prompt, it just directly recommends a few apps and strategies along with the instructions that the user’s wallet needs to execute them.

Usage example

A perfect case where AURA is valuable in multiple ways is this: imagine that a new web3 app launches a reward program and it wants to give a rewards boost to holders of a specific NFT, in order to reach out to this community.

However, said holders have no immediate way of knowing that they get this benefit from holding their NFT. The web3 app may not have direct channels to reach out to this community, especially if said community is exclusive.

In this case, if you’re a holder of this NFT and your wallet supports AURA, you can be pointed to this web3 app. This is beneficial on many levels:

• it's a specific strategy that you're unlikely to discover yourself
• it doesn't conflict with any other strategies, it supplements them
• it's a recommendation that actually boosts your portfolio, rather than simply being an ad (there’s benefit for the user)
• it's the best channel for the web3 app to communicate this, because it's falling into right the hands of users who are already looking for yield boost

The best example of this is LobsterDAO and Ambire Wallet, especially considering that this is a triple-dip opportunity for users (original asset, which could be yield-bearing, plus the Ambire rewards, plus the boost from the NFT).

Roadmap

v0.1: text-based recommendations (just runs security processing): at this stage, AURA will merely be an API that gets in an account address and outputs a list of app recommendations and strategies in natural language form, with a description of what each one does.

Following this stage, we’ll start an ongoing training of our own AI model behind AURA.

v0.2 basic external data sources: At this stage, AURA will pull in web3 apps, DeFi earning opportunities and airdrop opportunities from external sources, so as not to rely only on the information that the LLM model contains within it.