Architecture

Initial Stages

In the initial stages you will be able to watch the agent trade on DEXs and report its' daily PNL (profit and loss). You will also be able to connect your wallet and have the agent trade on your behalf based on your risk profile. This will have to be limited to holders of the token (beyond an amount to be determined). The rest of the community will benefit from portion of the fees paid in tokens being burnt.

Fully Autonomous Agent

While this is running we will be building the fully autonomous AI agent which will have the following layers:

• Decision making

• Blockchain Interaction

• Treasury

• Management

• Governance

Decision Making

Houses the trading strategies (both ML-based and rule-based). This will have the core AI trading engine. It will monitor real-time market conditions and decide which pairs to target, how to manage risk and rebalancing.

This will also have risk management strategies such as risk reduction, stop losses and fat-finger protections (to stop it from accidentally buying or selling too much at once). The agent will constantly back test its strategies to detect inefficiencies and propose changes which will be managed by the Governance layer.

Blockchain Interaction

Includes the wallet management and DEX integration (e.g. Raydium) using the relevant SDKs. Will be able to execute the trades after being instructed by the decision making layer. Makes sure that the data is clean and accurate before being sent to the decision making layer.

When the bot is fully autonomous it will be able to generate its own wallet addresses and manage the private keys. Once that wallet is seeded externally with funds only the agent will be able to control those funds for trading, use profits to buy GOODAI token and move to the treasury which will keep a portion and burn the rest.

In the earlier stages this layer will be connecting to an external user wallet to place trades on their behalf.

Treasury

This layer will receive the trading profits and decide how much to burn vs accumulate for funding operation. It will also be the accounting layer that will verify the performance of the agent and produce accounts.

Management

If the agent is to be fully autonomous it will need to be able to manage its own hosting and development fees. This raises some legal and ethical questions around accountability if the agent starts acting maliciously or causes harm especially since the agent will have funds that it controls. Even if these payments happen automatically paid in a certain crypto, it will still depend on physical resources such as a data center or cloud provider can decide to terminate the service if the agents activities are flagged or violating terms of service.

Governance

This layer will allow the agent itself or the community to propose updates, feature requests or change the direction of strategy. It will use SPL Governance to create a DAO.

Workflow example

• Agent deployed and generates a wallet

• good.ai as an external entity will seed the wallet with some SOL, stable coins and a collection of other tokens.

• The agent analyzes and sees an opportunity on Raydium for a tradeable launched token

• It swaps SOL for the token

• Over time it exists its position in a profit

• On realizing this profit it will use a portion to buy GOODAI

• Immediately it will burn a pre-defined portion

• The rest is put in the treasury account