Gmeow! Last week, we hosted our first AMA on our Telegram Community Channel , and in this article, we collected all the answers to it.
Enjoy, cats!
Q—So, Replicats is fully dependent on data. How accurate would the agent be in terms of some crazy market moves and sentiments? I saw the team not just building the regular agents, so it makes me curious.
[@quantamentalguy] A—Regarding market movements (structural market breaks), we use a mix of multivariate signaling (i.e., from several sources such as high NVT momentum, small-caps, ETF movements, traditional asset integration (FX, indexes, econometric, etc.) and market microstructure factors.
Q - What are the token ownership requirements for accessing the services Replicats will provide?
[@bernardorq] A—We are discussing the details of the tokenomics. The token will be useful on the copy trading platform, and we're working on staking/gating details, among other incentives and buyback mechanisms.
Q- On my end, what makes Replicats different from other trading agents? Who are your biggest competitors?
[@bernardorq] We're following other players, to quote a few: Asym, Almanak, HeyAnon, Griffain, Boltrade, and Croy. Some are building DeFi tooling, and others are building more autonomous trading. We want to work together with the tooling ones that can improve our execution.
We want to have autonomous agents on our Marketplace so our users can copy-trade them. There are many ways to partner and explore synergies.
@DegenKamp Q -Can you shed more light on the two specialized foundation models (time series and graph-based)?
[ @lucasfonsecaeng] Regarding temporal analysis, the GNN that manages the portfolio will input historical market data and trade signaling based on this data (with TA and Quant approaches). Another signaling scope is precisely the metrics related to the network topology (signaling of market microstructure factors).
A variant of factor investing reduces the complexity of all this data, classifying these signals by a maximum explanation of viability.
@kantacosmos Q—This might be dumb, but how will this agent work during trades? Do I deposit and leave the agent to buy whatever, or do I still have to instruct the agent literally?
The agent will come with customizable risk management parameters, allowing you to choose the best fit for your profile. No full configuration is needed for developed agents. You'll use your non-custodial wallet on our platform to manage funds and set general parameters, such as desired portfolio segments (e.g., blue chips, meme coins, AI tokens, etc.)
@jumani232 Q - How exactly does the DAG model differ from LLM, and how is the team planning to implement that in the agent
The agent works on time series prediction (GNN) and MPT-based portfolio management parameters. In this context, many parameterizations can change its behavior. The LLM comes in as a layer to translate the agent's multidimensional insights into a human form and receive these parameterizations.
Q - Would there be any major CEX listing as we progress?
[@bernardorq] We are speaking with some tier-2 exchanges and want to list only when the retail users from the CEXs can see Replicat-One running and copy-trade it. We have good connections with tier-1 exchanges but don't want to spend energy there yet.
Q- What major partnership is the team working on?
[@bernardorq] We're speaking with chains (Berachain, Polygon, Avalanche, Arbitrum, Mode, Ripple, Solana, etc.) and other AI Agents that focus on signals and predictions to explore building trading agents based on their predictions using our portfolio and risk management framework. Some KOLs/Traders are also taking a similar approach, building agents with their thesis.
This week, we had a call with a trading desk from a bank interested in understanding how they could use our framework. We're also speaking with dozens of data sources to improve our data lakes. We started conversations with DeFi protocols like Euler and Morpho to better understand how to implement strategies on them.
Q - What is happening now on Replicats, and what is coming?
[CTO] We're working hard to launch our first on-chain trading agent, built around our thesis. We anticipate having our data integration module, portfolio and risk engine module, and execution trading module ready.
@galadima Q - What other notable functions can the agent perform besides automating trades?
Autonomous Trading Agents: Self-learning Agents that execute trades in 24/7/365.
Sophisticated Risk Management: Agents aren't just hunting for quick profits. They are wired to focus on risk management, aiming for consistent upsides while minimizing potential blow-ups.
Blockchain abstraction: empowering users with social logins, non-custodial wallets, and abstracting concepts like hashes and block confirmations. Blockchains are the backend, not the product.
Deep Market Insight: By combining advanced time series analysis and graph-based modeling, Agents see hidden patterns and relationships users would likely miss. Agents get more reliable insights because they're not just skimming price data but decoding the market's structural heartbeat.
@kantacosmos Q - What's your distinct definition of AiFi and DeFAi? Saw an article from one for the team, and it was quite worth the read. (Bernardo)
[Bernardo] Thanks! We like AIFi because it's a broader vision in which AI transforms the entire landscape of finance, leveraging blockchain, DeFi, and RWAs. This makes it inclusive and easy to use for those outside the traditional financial and crypto ecosystems.
You can read more in this article
DeFAI is an interesting crypto narrative that caters to a small audience. We want to broaden it to serve a crowd not yet catered to by banks and who are not crypto natives. OFC crypto natives and degens are more than welcome!
@kantacosmos Q—Lastly, what market or chain would the agent be able to trade on? Do you have any plans to add others in the future?
We'll initially focus on EVM chains, ensuring a solid foundation, and will integrate Solana as soon as possible to expand our capabilities.
@yung_dnzel Q - What is the reasoning behind the specific portfolio composition of Replicat-One, and how does it align with your vision?
[Bernardo] The crypto market-tuned efficient frontier allocation on onchain books is a new feature we are working on: how to provide optimal sharp/vol on the very particular vol of the cryptos markets.
Replicat-One will manage a Base Ecosystem Long-only Momentum portfolio.
In short, a long-only actively managed portfolio of BTC (cbBTC), ETH, DeFi tokens (like Morpho, Pendle, and Aerodrome), AI (like Virtuals, Aixbt, and Vader), and Memecoins (Brett, Toshi) on Base.
The inspiration comes from the challenge of beating Bitcoin with controlled risk during this bull run. It's really hard for the average investor to decide when and how to rebalance all the different assets, navigate narratives and trends that change every other day, and manage volatility and risk properly.
So, the goal is to generate total returns greater than holding BTC alone with controlled volatility and better Sharpe by actively rebalancing the portfolio. Smart exposure to DeFi, AI, and Memecoins while not losing track of the Bitcoin general trend (tracking error with balanced alpha).
We hope that you all enjoyed and met us at our first X Spaces (announcement soon!)
Don't forget to join our community at
Till next time!
Replicat-ONE
