we launched the first ever fully autonamous trading agent in February, 2024. now, we're building the framework to empower others to do the same.
Muriel "Mickie" Seibert was the first woman to become a trader at the New York Stock Exchange.
In 1967, she changed the rules of trading.
Today, Mickie.AI is doing the same.
Mickie.AI is a framework for building fully autonomous trading agents; extending the core ElizaOS agent framework to enable the work streams, thought processes, and data access points needed for intelligent trading.
In the next years, we know that agents will handle most if not all aspects of trading on behalf of investors. We're here to make that timeline as short as possible by fueling a Cambrian Explosion of automatous trading agents - iterating and improving together.
In February 2025 our team created the first ever autonomous trading agent during an AI agent Hackathon. We wanted to answer the question: why has no one launched a fully autonomous trading agent? There are so many teams that are claiming to be "working on" autonomous trading agents; raising funds, launching tokens, and doing marketing for their products. But, none of them have launched a real agent. Could it really be so hard?
On Friday February 7th it went live.
In it's first week, our agent completed over 6000 trades on mainnet in a matter of days, and made some trades that had over 400% profit.
But, through launching the first autonomous trading agent, we also learned why no one had done so before.
There is a big innovation gap in autonomous trading agents, and it's for a good reason.
It is hard to make autonomous trading agents profitable.
With our agent, even though she makes some trades that are highly profitable, in the end she's still loosing money. It will be a long process of iteration to get to the point of where agents are profitable - let alone able to compete with the other option on the market for automatized trading: trading bots.
Compared to trading bots, which just automate complex mathematical algorithms to place profitable trades, autonomous trading agents are totally inferior today.
Despite this, we believe that eventually autonomous trading agents will be able to compete with algorithmic trading bots: in between today and that day, we're facing a BIG innovation gap.
What will get us across this chasm?
Iteration and knowledge sharing.
The more teams building autonomous trading agents, sharing each other's learnings, and iterating off each other's code the better. The more we iterate, the faster we will get to the day where agent's aren't just profitable, but are capable of powering financial market efficiency at levels we've never seen before.
The future of autonomous agents depends on this.
That's why we're taking action.
No one wants to launch an autonomous agent that is not profitable.
But, if no one is launching these agents publicly, iteration is doomed to be slow and inefficient.
This fear to launch agents that aren't yet profitable is compounded by many of the "teams working on autonomous agents" already having raised VC capital and launched a token.
If they deploy and share an agent that is not profitable, then this is something that could have dramatic negative effects on their token price and appeal to venture capital.
In the earliest days, all autonomous trading agents will be highly unprofitable.
We need a way to incentivize the development of these unprofitable agents, until the point where innovation breaks the barrier of profitability.
The future of autonomous agents depends on this.
It's our mission to build a framework and coalition of teams working to help autonomous agents cross the innovation gap.
We are working on:
Sound like something you want to be a part of? DM Kirsten on Telegram.
The Mickie.AI Framework is currently in development and evolving rapidly.
We aim to publish a first version near the end of February 2024.
In the meantime, we have public information on our original fully autonomous trading agent. This agent is live on Solana mainnet and making trades today. We are basing the framework off of this agent, and therefore it's a great starting point for teams that want to start shipping NOW!
It's 100% open-source and can be forked and iterated upon today!
Our origional autonamous trading agent was under the brand CookFi AI during a hackathon. We're in the process of merging the CookFi codebase with Elyrai AI, another agent we maintain. The Elyrai AI agent which will be our main example agent using the Mickie.AI Framework going forward. As we extend the framework and add new partner technologies, we will continue updating Elyrai AI and using it as a case study for other teams to leverage.
We're a developer collective that's been building products together for years. Now, we're growing the movement behind Mickie.AI.
We believe we're working on the most exciting emergent area of AI agent development.
If you believe the same, join us.
DM Kirsten on Telegram to get looped in and join the developer collective.