This article is part two in my series of exploring the Cod3x ecosystem and the capabilities of its AI agent platform.
If you haven’t read the first article in the series on setting up your agent with an advanced “Trading Strategy” I suggest you give that a read first.
After setting up your agent, you will be ready to explore this article where we learn how to “Pair Trade” with our agent.
For the uninformed, “Pair Trading” is similar to “Pair Coding” from the AI realm where software engineers work in tandem with AI to complete tasks in code.
This can look like a conversation between the engineer and AI, or something as simple as asking the AI for code recommendations the engineer will then manually act upon.
Some engineers take this idea so far as to simply direct the AI on the intent of their outcome and guide the Agent while it writes code itself.
In this article, we will be exploring the idea of having our Agent provide us with trading positions we can choose to manually act upon.
Chatting with your Agent
In order to understand pair trading with our Agent and their strategy, we first need to understand how the Cod3x chat works and interacts with the plugins the Cod3x team has built.
1. Navigating to the chat interface
The chat interface can be found by navigating to your agent by clicking its name in the side nav seen below that we discussed in the previous article.
Clicking the first button in the side nav will expand it and allow you to select the agent you wish to chat with.
After this, you should see the Cod3x Trading Terminal with the chat interface on the right hand side of the screen
In order to chat with our agent, we will need to “Activate Agent”.
2. Activating the agent
Side note:
At the time of writing this article you will need to hold 100,00 CDX tokens, or apply for the open beta in order to activate your agent. As time goes on the team will lower the requirement as they scale up their system and push the capabilities out of beta testing which they are currently in.
If you want to know when this requirement will be lowered, or want to apply for the beta testing white list, join the discord or pay attention to the Cod3x twitter for announcements.
To activate our agent, we will need to click the settings icon in the top right of the page above the chat interface.
Here you will find the ability to activate your agent which will turn it on and allow it to chat or execute goals.
After we activate our agent, the agent status will flip from red to green as seen with my already active agent named Auriana in the image below.
For the rest of this tutorial we will be using Auriana given she is active and meets the activation requirements.
You can see that I’ve activated her and claimed the “Early Access Credits” available to those with 100,000 CDX beta testing the platform.
3. Purchasing Credits
In order to chat with our agent we will need “Credits” from the Cod3x platform.
10,000 credits can be claimed for early access users at the time of writing this article, or they can be purchased through Coinbase commerce.
To purchase credits, we first need to navigate to our wallet which can be found in the lower left hand corner of the screen
After clicking “Buy Credits” you will be prompted with the packages available.
At the time of writing and testing this article, I have found the 11,000 credit pack to be solid for a week of analysis when analyzing for positions once daily while the stock exchange is open, on 5 separate coins with the current strategy in place.
The pricing may seem high, but given we are using the credits to earn money, they will pay for themselves over time depending on the size and win rate of your personal agent.
After clicking the button to “Purchase with Coinbase” you will be given a “Transaction ID” and the ability to “Pay Now” for the credits with a linked wallet on the Coinbase commerce page
Clicking the link to “Pay Now” brings us to the Coinbase commerce page, where you can connect the wallet of your choosing to make the payment.
Warning: at the time of writing this article, clicking “Later” will keep the transaction open with Coinbase commerce, but there is no way to view the transaction request from the trading terminal. In a future update, you will be able to find the pending transaction in the alerts section or other area of the app. If you end up choosing “Later” and lose the page, you can always initiate a new, fresh transaction.
After completing your payment, you will receive a receipt. Keep this somewhere safe in case the payment doesn’t go through, and you need proof for the Cod3x team to investigate.
Now that we have some credits visible in our wallet, we can begin chatting with our agent and creating goals.
To view the credits you currently have, you can refresh the page and click on the wallet icon in the bottom left of the screen.
The chat allows you to see how many credits your agent is using in its active thought process, along with the number of tools used. This can be insightful when attempting to optimize for credit usage (which we will be doing in a later article)
4. Chatting with the agent
Chatting with the Cod3x agent we’ve built in this article series is very similar to the ChatGPT 5 interface.
There is a text input at the bottom of the interface with the ability to enter any user message you desire.
There is also the ability to select the model of your choice in the top right of the chat interface.
As I mentioned in the previous article, Claude Sonnet 4 is the best model for financial analysis, although it may be more expensive per chat compared to a smaller model like ChatGPT’s 04-mini, it is far more performant for our use.
Recently Claude has released Claude Sonnet 4.5 which we will default to once it is available in Cod3x.
At the time of writing this article 4.5 was not available, but has recently been added
When we interact with the chat, it will automatically inject our “Trading Strategy” before the message is sent to give the AI context on how we want our agent to behave
If a user sends the following message
Cod3x will perform work in the background to inject context similar to the example below.
Trading Strategy: (this is where your strategy is inserted)
Personality: Balanced, etc
User Message: Hi, I would like to open a long positionYou as the user will not see any of this context they inject, but the tooling Cod3x has built will automatically inject the necessary context to complete the chat request.
5. Crafting a message to send to our agent
The key to getting maximum value extracted from the markets by your agent is to carefully craft your chat messages and interactions.