# Replicats

Smart Wealth Companions

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FGBvvjYZlC9h2ogEExLZT%2FCOVER%20X%201500x500.png?alt=media&amp;token=fc4feed6-fce6-45d1-8e7b-47f23fb08a62" alt=""><figcaption></figcaption></figure>

## Replicats - Smart Wealth Companions

Welcome to Replicats, your first portfolio on autopilot.

Replicats is an AI platform that offers Smart Wealth Companions. These are agents designed to grow and protect your wealth using institutional-grade strategies, while you focus on everything else. Launch your agent, set your risk profile, and let it execute 24/7.&#x20;

You maintain full custody. Your agent handles the rest.

## **Smart Wealth Companions for the 99%**

Since January 2025, we've been building the infrastructure for autonomous portfolio execution. Our flagship agent, Replicat-ONE, has been trading live on Base since February 2025, actively managing portfolios of **BTC**, **stETH**, **Solana** (on Base), and **Morpho** using proprietary quantitative frameworks.

The problem is simple: systematic portfolio management requires continuous discipline that human traders can't sustain. Market volatility rewards systematic execution but punishes emotional reactions.&#x20;

Most investors understand the theory but lack the infrastructure to implement professional-grade risk management around the clock. Replicats provides that infrastructure.<br>

To start your journey, sign up now at [www.replicats.ai](https://www.replicats.ai)

***


# Smart Wealth Companions

You can easily launch your own Replicat-ONE agent with just a few clicks and a minimum investment of $50 (USDC). Replicat-ONE makes trades based on quantitative signals, rebalances positions in response to market changes, and enforces institutional risk controls, including **Value-at-Risk (VaR)** and **Conditional Value-at-Risk (CVaR)**.&#x20;

Each agent is customized to match your risk profile (Conservative, Moderate, or Aggressive). All trading is conducted on-chain with full transparency, and your funds are accessible only to you.&#x20;

Replicat-ONE does not support reactive signal-chasing or discretionary trading. Instead, it uses systematic portfolio building, factor-based selection, and adaptive risk management, supported by professional quantitative methods, available through a monthly subscription.&#x20;

Replicats plans to introduce additional agents in the future that will enable access to various assets, including stocks, the S\&P 500, gold, and RWAs.


# Mission and Values

## Our Mission

Our mission is to **make sophisticated investment management accessible, intuitive, and secure for everyone, whether seasoned crypto traders or new investors**.

**Replicats** bridges these gaps with AI investment agents that autonomously manage portfolios, learn dynamically from the markets, and communicate transparently through conversational interfaces.

Many people globally lack access to effective and high-quality investment services. Crypto markets, despite their potential, can be volatile and challenging to navigate. Traditional solutions often fail by being overly complex, inaccessible, or insufficient in managing risks.

***

## Our Building Values

**Replicats** employs advanced representation learning and adaptive risk management techniques using multi-modal AI to deliver:

1. **Agents Built Specifically for Trading** — We use specialized models like Graph Neural Networks and time-series forecasting engines designed specifically for financial markets, ensuring precise decision-making and trade execution without the unpredictability of large language models.
2. **Global Accessibility** — Investment opportunities shouldn't depend on where you were born or how much wealth you started with. We break down barriers to financial growth worldwide, using blockchain as rails.
3. **Intelligent Simplicity** — We make sophisticated investment management accessible to everyone without requiring technical expertise or sacrificing performance. There are no black boxes or hidden agendas.
4. **Capital Preservation First** — We believe growing wealth means nothing if you can't preserve it. Every decision prioritizes protecting what our users have earned, so we focus on institutional-grade risk management for our agents.
5. **Adaptive Tech** — We don't impose one-size-fits-all strategies. Our agents learn your risk tolerance and adapt to your goals, working as your personal investment partner. Chat terminal available!


# Target Users

**Replicats** serves three distinct audiences, each with unique needs but a shared desire for intelligent investment management.&#x20;

From crypto natives seeking to preserve their gains to traditional investors exploring digital assets, our AI agents adapt to different risk profiles and investment goals while maintaining the same core promise: professional-grade portfolio management that protects your capital.

* For Retail Investors looking to grow their wealth
* For Agent Builders interested in generating revenue by sharing strategies
* For Web3 Wallets and Exchanges that want to offer Replicats to their users.


# For Smart Investors

We're building for the billions left behind by traditional finance. Whether you're deeply embedded in crypto, checking your wallet every month, or just getting started, our goal is the same: to give you an AI investment partner that works with your habits, your risk tolerance, and your life.&#x20;

* **Crypto Natives**: Tech-savvy investors looking for advanced tools to optimize portfolio management.
* **Sporadic Investors**: Individuals with crypto holdings seeking a reliable way to manage risk without constant trading engagement.
* **Underserved Retail Investors**: Users from emerging economies require straightforward, secure, and effective access to global financial markets, including cryptocurrencies, stocks, and indexes.


# For Agent Builders

We're also creating space for professionals to share their expertise on a larger scale.&#x20;

Through our **Agent Launchpad**, experienced investors and strategists can use our framework to build, release, and monetize their AI investment agents.

To create Agents and strategies, we invite professionals such as:

* **Seasoned traders** with a proven track record who want to turn their strategy into an Agent.
* **Hedge fund managers** who are looking to deploy algorithmic strategies at scale.
* **Quant engineers and data scientists** who want to experiment with multi-modal models in a real-market environment.


# For Web3 Wallets, dApps and Exchanges

Partner with Replicats to transform your platform into a comprehensive investment ecosystem. Our white-label AI investment agents help you retain users, generate recurring revenue, and differentiate from basic trading tools—all while offering your users institutional-grade portfolio management.

Some of the potential integrators are:

* **Centralized Exchanges:** Generate new revenue streams beyond trading fees by offering AI-powered portfolio management to your user base.
* **Web3 Wallets:** Keep users engaged longer with set-and-forget investment solutions that work across crypto and traditional markets.
* **Trading Applications:** Stand out from competitors by offering intelligent portfolio management instead of just charting and analysis tools.


# Team

Who's behind Replicats.

<table><thead><tr><th width="149.34375">Who</th><th width="183.24609375">Name, Role</th><th width="418.125">Bio</th></tr></thead><tbody><tr><td><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FaVvPvSNV981BnNQWQgoS%2FBernardo-Quintao-AAA-1.jpg?alt=media&amp;token=8562dbcc-9678-420b-a646-8d3ef663ca28" alt=""></td><td><strong>Bernardo Quintão</strong> CEO, BD &#x26; Strategy</td><td><ul><li>Crypto for 10y+.</li><li>Former Head of BD @ <strong>Backed.fi</strong>.</li><li>Former Head of International BD @ <strong>Mercado Bitcoin</strong>.</li><li>Past life in TradFi, M&#x26;A, Sales &#x26; Trading. Serial Entrepreneur.</li><li><a href="https://www.linkedin.com/in/bquintao/">Linkedin</a>, <a href="https://x.com/bernardorq">X</a>.</li></ul></td></tr><tr><td><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2Fn16hYcUX7IC0fhXPlayC%2FMagdalena%20Pire.jpg?alt=media&amp;token=c5df7757-520b-436f-be09-409aab6570e0" alt=""></td><td><p><strong>Magdalena Pire Schmidt</strong> <br></p><p>COO</p></td><td><p></p><ul><li>MA in Linguistics.</li><li>15 years in tech ops and consulting. Led ops teams in <strong>Google</strong> and <strong>Adyen</strong>.</li><li><a href="https://www.linkedin.com/in/magdalenapireschmidt/">LinkedIn</a></li></ul></td></tr><tr><td><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FA6DwDCZTYhojCf8U7CIq%2FAnon%20Dev.jpg?alt=media&amp;token=ed72e9fa-42bb-4c93-b182-9b1d23a5c6bf" alt=""></td><td><strong>Anon CTO</strong></td><td><p></p><ul><li>Crypto OG and experienced CTO.</li><li>Managed 200+ people teams in banks and crypto startups for 15+ years.</li><li>Built bridges and staking protocols across different blockchains.</li></ul></td></tr><tr><td><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FwjfcOXsy3nWwwfpCwPJw%2FBruno%20Alano.jpg?alt=media&amp;token=5bec183e-ccff-4cf0-ba76-c894d57573e2" alt=""></td><td><strong>Bruno Alano</strong><br><br>AI Advisor</td><td><p></p><ul><li>CTO &#x26; Co-founder @ <strong>Avra</strong></li><li>Early researcher at <strong>OpenAI</strong> (2016).</li><li>Youngest and first LATAM researcher @ OpenAI.</li><li>First AI/ML startup in 2014.</li><li><a href="https://www.linkedin.com/in/brunoalano/">LinkedIn</a>, <a href="https://x.com/brunoalano">X</a>.</li></ul></td></tr><tr><td><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FWsczrebzmYkQIf9QJ6ZE%2FScreenshot%202025-06-05%20at%2015.09.30.png?alt=media&amp;token=cf6afe60-a214-4916-81b8-fc4d1574e01f" alt=""></td><td><p><strong>Guilherme Reis</strong> <br></p><p>Head of Quant</p></td><td><ul><li>Quantitative researcher and algorithmic trading specialist with expertise in machine learning-driven investment strategies. </li><li>Served as Director at Scotiabank Canada.</li><li>Former Vice-President at Citco.</li><li> MBA in Finance, an LL.M. in Finance, and is a CQF Institute-certified quantitative finance professional.</li><li><a href="https://www.linkedin.com/in/guibertoni/">Linkedin</a></li></ul></td></tr><tr><td><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FVXmNOQIOPYUjeTVSRt4Z%2FBreno-Melo.jpg?alt=media&amp;token=5ab248ec-dca1-4cb9-96f6-9ccf121e24b7" alt=""></td><td><strong>Breno Mello</strong><br><br>Quant Researcher</td><td><ul><li>CQF-certified quant researcher/engineer specializing in deep learning and feature engineering. BSc in Economics</li><li>Experienced in quantitative strategies for institutional clients, portfolio optimization, and advanced risk management.</li><li>Proven track record in derivatives pricing and options strategies.</li><li><a href="https://www.linkedin.com/in/brenodemelo/">LinkedIn</a></li></ul></td></tr><tr><td><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FSvXzpy3INBpm6yLZEsbe%2FGiorgio-Giuliani.jpg?alt=media&amp;token=e004bae0-c555-4ace-8abb-ed3f2560869f" alt=""></td><td><p><strong>Giorgio Giuliani</strong></p><p></p><p>Head of Product</p></td><td><ul><li>Engineer working in product management for over 10 years</li><li>Specialized in products and applications of different types, from core banking systems to lending pipelines to Financial API and tokenization engines.</li><li>Formerly <strong>Backed.Finance</strong>, <strong>Funding Circle</strong>, <strong>Sumup</strong>, and <strong>N26.</strong></li><li><a href="https://it.linkedin.com/in/giorgiogiuliani">Linkedin</a>, <a href="https://x.com/ggglni">X</a>.</li></ul></td></tr><tr><td><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FONWIzd5XWgyNpDRe3FsT%2FSamuel-Eidam.jpg?alt=media&amp;token=905eb884-fc32-491d-a005-c4d3fc2fd2b1" alt=""></td><td><strong>Samuel Eidam</strong><br><br>Product Designer</td><td>TBD</td></tr><tr><td><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FcHZ2e7IUypRZKyZYAjlA%2FGuto%20Martino.jpg?alt=media&amp;token=ae76605e-adc0-420d-994a-9bc1e0de4fd7" alt=""></td><td><p><strong>Guto Martino</strong> <br></p><p>Head of Marketing</p></td><td><p></p><ul><li>12+ Years in Marketing. <strong>Safary Club Member</strong>.</li><li>Founding member of <strong>Ethereum Rio</strong> and <strong>Department of Decentralization/Berlin</strong>.</li><li>Former Head of Marketing at <strong>Astaria</strong> and  <strong>TokenSight</strong>, <strong>Near Protocol</strong>, <strong>Hathor</strong>, and <strong>Arkis</strong>.</li><li><a href="https://www.linkedin.com/in/brunoalano/">Linkedin</a>, <a href="https://x.com/gutomartino">X</a>.</li></ul></td></tr><tr><td><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FSwAsC0VFsOchVsgXHofU%2FJosep%20Kingsley.jpg?alt=media&amp;token=d64d78c2-96bc-498f-a440-286f1bff94aa" alt=""></td><td><strong>Josep Kingsley</strong><br><br>Community Lead</td><td><p></p><ul><li>Blockchain Researcher 7+ years in crypto</li><li>Community Growth Strategist</li><li>Previously <strong>Tokensight</strong>, <strong>SmartFox</strong></li></ul></td></tr></tbody></table>


# Business Model

{% hint style="info" %}
**We seek to explore different ways for the platform to generate revenue.**

**IMPORTANT**: Replicats has no governance or utility token yet. Our first launch is a token for the Replicat-One agent.
{% endhint %}

**Possible revenue** **streams**:

* New Agent fee:
  * A fixed fee paid in advance. The fee amount depends on how often the **Agent** will analyze the market and its own portfolio.
* Maintenance fee:
  * A monthly fee is deducted from the Agent's wallet. The fee amount depends on how often the Agent will analyze the market and its own portfolio.
* Premium features:
  * Feature-specific one-time fees. E.g., the ability to manually trigger a market check.


# Roadmap & Sprints

"Go instead where there is no path and leave a trail". - Ralph Waldo Emerson

{% stepper %}
{% step %}

### Q1 2025

*In our first quarter, the team formed and we built the foundations of the Replicats Intelligence Framework.*<br>

* Sign ups and wallets.
* Deposits in EVM and Solana on Terminal.
* Replicat-One Token launch.
* First agent POC live.
* Full-time team.
  {% endstep %}

{% step %}

### Q2 2025

*Early users start benefiting from the power of Replicats Intelligence Framework by copying Replicat-one into their wallets; Framework becomes more reliable. The team explores additional assets and starts scaling.*

* Replicat-One Agent trading live. Users can make copy into their wallet.
* Improve reliability as measure as downtime.
* Complete legal setup.
* Validate pricing.
* Integrate 0g.
  {% endstep %}

{% step %}

### Q3 2025

*Users will be able to trade additional assets and increase the customization of their agents.*&#x20;

* Launch new agents, including&#x20;
* More assets (e.g. bCSPX(S\&P500) and PAXG(Gold)), AI models and trading venues.
* Test new features and revenue lines.
  {% endstep %}

{% step %}

### Q4 2025

*We set up the framework to allow users to develop and share their own trading agents.*

* Develop and test Launchpad
* More AI models and data sources
* Enterprise customers
  {% endstep %}

{% step %}

### 2026

*Our users can choose agents from a rich marketplace of agents and assets.*

{% endstep %}
{% endstepper %}


# Sprint #1

**Timeframe**: Jan-Feb 2025

Please check the [End of Sprint #1](https://paragraph.com/@replicatsai/end-of-sprint-1) blog post for a more in-depth reading.

***

## Product Overview & Vision

**Replicats** is an AI-driven trading framework that goes beyond typical “Crypto AI Agents” by combining:

* **Predictive Analytics**: Using time-series models for price forecasting with quantiles.
* **Workflow Orchestration**: Trigger-based execution of agents (e.g., scheduled checks, on-chain data events).
* **Embedded Wallets**: Automated (or user-approved) buy/sell transactions on both **EVM** and **Solana** networks.
* **Copy-Trading**: Users can “fork” or copy-trade an existing Agent (e.g., “Replicat-ONE”) and view PNL in a dashboard. **\[Deprecated ❌]**

## Sprint 1

## Primary Goals of the v0 Prototype

1. **Implement a Trigger and workflow System: This is distinct** from the web dashboard. It enables Agent logic to fire automatically when certain conditions are met.
2. **Demonstrate Replicat-ONE**: A hard-coded strategy/Agent that uses forecast data and on-chain metrics to make trading decisions.
3. **Allow Copy-Trading**: Users can “fork” Replicat-ONE in the dashboard, letting them replicate trades in their own embedded wallets. \[Deprecated ❌]
4. **Show Trading History & PNL**: Provide a simple interface to view the original agent’s and the user’s forked agent’s transactions and performance.

## Key Components

1. **Data Ingestion & Database**
   * Regularly fetch raw market data from Data sources.
   * Store the data in a database.
   * Provide a **Pub/Sub or event** indicating data changes so agents can react (if relevant).
2. **Forecasting Module**
   * A separate service (likely containerized in Python) that runs **our models** every X minutes (e.g., hourly).
   * Reads the latest price data from the database, generates forecasts (median, 5–95% quantiles), and writes them back to a “forecasts” table.
   * Emits an event/message when new forecasts are available.
3. **Agent Execution Service (Workflows & Triggers)**
   * **Core**: A job or microservice that runs continuously, listening for triggers (changes in DB, new forecasts, scheduled intervals).
   * **Workflow Logic**:
     1. **Define triggers** (e.g., “If new coin on Solana has < $1M market cap but > 500 holders,” or “Run this Dune query every 15 min,” or “When forecast probability of 5× > 20%”).
     2. **Execute Agent logic** upon triggers (buy, sell, or more complex workflows).
   * **Replicat-ONE**: A special Agent (hard-coded) that trades based on the forecasting data.
   * **Copy-Trading**: **\[Deprecated ❌]**
     * If a user has chosen to copy-trade Replicat-ONE, the same actions are triggered on the user’s embedded wallet.
4. **Dashboard (Next.js + tRPC)**
   * **Authentication & Wallet Management**: Uses wallet infrastructure to create 2 embedded wallets (EVM & Solana). Users can also fund/withdraw.
   * **“Fork” or Copy-Trading UI**: A toggle/button to replicate Replicat-ONE.
   * **Transaction & PNL View**:
     * Show Replicat-ONE’s trade history and performance (PNL).
     * Show the user’s forked-agent transactions and PNL.
   * **Minimal Agent Customization (Future)**: For v0, we only allow forking Replicat-ONE. Later, we can allow custom triggers and logic.


# Sprint #2

**Timeframe**: Feb 2025 - March 2025

Please check the [End of Sprint #2 blog post](https://paragraph.com/@replicatsai/end-of-sprint-02) for a more in-depth reading.

***

## 1 - **Completed**

#### 1.2 - External Dashboard

* Operational dashboard for internal and external use.
* Onboarding of the first Early testers

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FNkYjIFXjpXaaJyYm1aG2%2Fimage%20(2)%20(1).png?alt=media&amp;token=ba8b37f0-3b80-4ba8-86db-4ffe7c6a5aa9" alt=""><figcaption></figcaption></figure>

#### **1.2 - Dashboards & Monitoring**

* Develop External Dashboard (Replicat One) for early adopters.
* Build a Data Pipeline Dashboard to monitor ingestion and ETL.
* Implement a Monitoring System:
  * Alerts for CPU/memory thresholds.
  * Checks for the latest data ingested.
* Expand Dashboard APIs (in coordination with the tech team).

#### **1.3 - Agent Improvements**

* Improve Agent performance and reliability.
* Refactor Agent execution logic by user (simplify if necessary).
* Create staging environments for:
  * Agent
  * Replicats API
  * Data pipeline

***

#### **1.4 - LLM Integration & Configurable Agents**

* Enable dynamic Agent configuration through LLM (RAG-based).
* Allow users to schedule future configuration changes via LLM.
* Add relevant RAG modules to support parameter customization and scheduling.

***

#### **1.5 - Data & Signal Development**

* Start CoinGecko data ingestion.
* Build an ETL pipeline to fill in missing or incorrect data.
* Conduct Hurst exponent analysis (market regime shift detection).
* Integrate Acolyte mindshare data for trading signals.
* Define the liquid index using Dune data.
* Use the CME volatility surface (PDF-based) for long signals.
* Add Granger causality to feature selection and calibration pipeline.

***

#### **1.6 - Documentation & Internal Enablement**

* Document models, architecture, and processes.
* Continue internal knowledge transfer:
  * Code reviews
  * Foundations of Agent models and implementation

***

#### **1.7 - Additional Features & Considerations**

* Implement a staking/gating mechanism for access control.
* Add support for time-limited and freemium access models.

***

## 2 - Spills for [Sprint #3 \[Current\]](/introduction/roadmap/sprint-3-current)

* Improving API’s performance
* Data layer restructure for the application
* System upgrade for Multiple Agent Deployment


# Sprint #3 \[Current]

**Timeframe**: April 2025 - May 2025

***

#### Core Infrastructure

* Refactor Agent architecture.&#x20;
* Create a new database.<br>

**Refactor APIs**

* Separate EOD and calibration processes.
* Separate forecast from Admin process
* Add JWT-based authentication to the API.
* Evaluate and integrate Dynamic's SDK into the API.

#### Pipelines & Terminals

* Develop a unified pipeline for Agent deployment.
* Improve terminal features
* Add forecast graph
* Build a Multiple Agent Terminal for centralized Agent management.
* Enable automated deployment of agents.

\
**Quant Features**

* Integration with Coingecko for data providing
* Integration for financial, politics, and market data with \[Redacted]
* Risk Mandate improvements for Replicat-ONE and the framework

<br>


# Agent Launchpad

Replicats offers Agent builders a crowdfunding platform (launchpad) to obtain funding and initiate personalized Agents. Later, other users can copy and deploy these Agents in their wallets, creating a positive flywheel effect.

Our mechanic

* Every new Agent owns a [trading wallet ](/old-technical-foundations/platform-architecture/wallet-system)and initial trading capital from the LP.
* If successful, these agents will be listed in the Replicats Marketplace.&#x20;
* Other users can clone them and deploy them to their wallets.
* Agent creators receive a fee.
* A token launch could also occur.

{% hint style="info" %}
We aim to launch the Agent Launchpad  in Q1 2026
{% endhint %}


# Getting Started

**Replicats** is an open playground for anyone looking to experiment with autonomous investing passively and actively.&#x20;

Whether you're a seasoned crypto trader, a curious tinkerer, or a professional hedge fund trader, we give you the infrastructure and framework to create, test, and evolve AI Agents that actually trade.

With **Replicats**, users can:

> **Step 1—** Create an account on [www.replicats.ai](http://www.replicats.ai) with an email or Web3 wallet. \
> [This will generate a unique non-custodial wallet](/old-technical-foundations/platform-architecture/wallet-system) for each user.
>
> **Step 2—**&#x46;und their non-custodial wallet using BTC, ETH, or stablecoins. We will also enable the fiat on-ramp in the future.
>
> **Step 3—** Assign one or more AI Agents to manage their wallets by:\
> &#x20;      a) Deploying a pre-built Agent from the Marketplace, or
>
> &#x20;      b) Creating a custom Agent using our no-code framework
>
> **Step 4—** Chat with your Agents via a natural-language terminal to fine-tune goals, risk levels, and asset preferences.
>
> **Step 5—** Get full transparency into trades, metrics, and Agent reasoning—no black boxes.

{% hint style="info" %}
Each Agent is built to learn and evolve with you. Whether you're backtesting a strategy or deploying real capital, Replicats gives you tools once reserved for hedge funds—now accessible through a conversational interface.
{% endhint %}


# Hiring!

We are actively looking for top talent to join Replicats. Apply now!

<table><thead><tr><th width="200.8828125">Position</th><th>Link</th></tr></thead><tbody><tr><td><strong>Quantitative Developer</strong></td><td><a href="https://www.linkedin.com/jobs/view/4232216622/">https://www.linkedin.com/jobs/view/4232216622/</a></td></tr></tbody></table>


# Getting Started

{% embed url="<https://youtu.be/PbfnED_dX2E>" %}

**Step 1** - Create a Replicats account using your email or Web3 wallet.

**Step 2** - Deposit USDC on Base. You can also buy USDC directly using a credit card through your Coinbase account.

**Step 3** - Select Replicat-ONE from available agents and launch it.

**Step 4** - Choose how much USDC you want to allocate to the agent.

**Step 5** - Select your risk profile (Conservative, Moderate, or Aggressive) and confirm your asset exposure preferences.

From that point forward, your agent operates autonomously 24/7.

{% hint style="info" %}
Disclaimer: Replicats is currently in beta. To get access, please contact us via email or Telegram.<br>
{% endhint %}


# Portfolio Monitoring

As soon as you launch it, Replicat-ONE will build your first portfolio, continuously analyze market conditions, and execute trades based on quantitative signals from our proprietary models.&#x20;

The agent rebalances your portfolio as market dynamics shift, adjusting position sizes to maintain optimal risk-adjusted exposure.&#x20;

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FOPmT3eRaWS0isgy7z6w9%2FScreenshot%202026-01-14%20at%2023.55.18.png?alt=media&amp;token=ef51ee73-2eaa-43df-a0b2-ae00347c6173" alt=""><figcaption></figcaption></figure>

Our dashboard provides a comprehensive view of your agent's activity and status, including:

* **Current portfolio:** the assets that your agent currently holds.
* **Past Transactions:** all the USDC deposits made into the agent portfolio
* **Past Orders:** What did your agent trade
* **Settings:** The associated risk measurements and maximum asset exposure

You can chat with the agent through an LLM terminal to understand:

* The reasoning behind each trade.
* The current market price of tradable assets.
* The agent's price forecast.


# Chat Terminal

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2F52UOjsetFkn7dpXBNFR2%2FScreenshot%202026-01-18%20at%2008.31.04.png?alt=media&amp;token=c8306372-c97c-497b-9448-249354acd491" alt=""><figcaption></figcaption></figure>

You can chat with the agent through an LLM terminal to understand:

* The reasoning behind each trade.
* The current market price of tradable assets.
* The agent's price forecast.

{% hint style="info" %}
We are actively developing the chat terminal. Some functionalities will become more stable with time.
{% endhint %}


# Risk Profile

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FZyUW5T3EAmXCfJr79DRH%2FG-FTeJBXoAAgcvf-1.jpg?alt=media&amp;token=f1427bf3-8a67-4784-9371-013935b621d0" alt=""><figcaption></figcaption></figure>

Your risk profile determines how aggressively your agent pursues returns versus capital preservation:

* **Conservative**: High USDC concentration. Lower volatility tolerance, tighter stop-losses, smaller position sizes. Prioritizes capital preservation while targeting modest growth.
* **Moderate**: The standard setup. Balanced approach between growth and protection. Standard risk parameters are suitable for most users.
* **Aggressive**: More USDC is allocated to the tradable asset. Higher tolerance for volatility, larger position sizes, wider exposure bands. Maximizes growth potential by accepting larger drawdowns due to allocation size.

The underlying strategy remains consistent across profiles, while your selection calibrates the amount of risk the agent takes in pursuit of returns.

<br>


# Maximum Asset Exposure

Currently, Replicat-ONE trades cbBTC, stETH, SOL (on Base), and MORPH (Morpho). All execution occurs on-chain through Base, making every trade verifiable and transparent.

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FT710DUpmo4rLWRKn2hYM%2FScreenshot%202026-01-14%20at%2023.46.47.png?alt=media&amp;token=11aa68ef-81e7-4467-9de1-5816e9b14a8d" alt=""><figcaption></figcaption></figure>

You can easily specify the minimum and maximum allocation for each asset with a few clicks. In case you prefer not to include an asset, you can simply toggle it off and the agent will not trade it.

{% hint style="info" %}
Important: your agent will always maintain 1% of the wallet allocated in $RCAT and will deduct 1 (one) $RCAT every 10 minutes to keep it running.
{% endhint %}


# Adding more Funds

You can add more USDC at any time, and the agent can manage it with only a few clicks.

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FfLzdSyrTjhA50IzEut2i%2FScreenshot%202026-01-15%20at%2000.08.57.png?alt=media&amp;token=46c8938c-009c-40e0-8b20-a0553381e515" alt=""><figcaption></figcaption></figure>


# Deactivating your agent

You can also deactivate the agent at any time, and your portfolio will be converted back to USDC and returned to your wallet immediately. No lock-ups, no withdrawal delays.

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FD08bql14rUbdCqbDXrvB%2Fezgif-2a354a2f5f3b1adf.gif?alt=media&amp;token=6e42725d-aa17-474a-ae91-440861b82b08" alt=""><figcaption></figcaption></figure>

By confirming deactivation, your agent will:

* Sell all holding assets back to USDC
* Transfer the USDC back from the agent wallet to your Available Funds

{% hint style="info" %}
Important: your entire trade history will be lost.
{% endhint %}


# Trading Mechanics

Here's the breakdown on how the agent makes its decisions:

* **Trend Analysis Across Timeframes: Replicat-ONE** tracks short- and long-term trends across multiple timeframes to determine when to enter or exit positions. It uses probability, not gut feeling.
* R**isk Management That Adapts**: The agent detects market regimes—trending, choppy, or crashing—and adjusts position sizes accordingly. When risk is high, exposure is low. When the signal is clear, it leans in.
* **On-Chain Execution:** The agent trades autonomously on Base, interacting directly with smart contracts for fast, cost-effective execution.

### What Replicat-ONE Trades

Currently, the agent builds a portfolio comprising cbBTC, stETH, SOL (Solana on Base), MORPH (Morpho), and USDC. These are liquid assets with established track records, not speculative low-caps. It only enters positions when the math favors the odds:

* **Momentum Signals** – Identifies assets gaining strength across timeframes
* **Risk-Adjusted Allocation** – Every position must improve the portfolio's risk-return balance
* **Volatility Awareness** – Scales exposure based on market conditions

{% hint style="info" %}
Important: your agent will always maintain 1% of the wallet allocated in $RCAT and will deduct 1 (one) $RCAT every 10 minutes to keep it running.
{% endhint %}

<br>


# Framework


# Onchain Trading


# Smart Wallet Custody


# SaaS Business Model

Replicats operates as Software-as-a-Service. We charge a monthly subscription for the technology, paid using $RCAT directly from the user's wallet.

Like OpenAI and Anthropic, Replicats charges users with tokens.&#x20;

{% hint style="info" %}
Important: your agent will always maintain 1% of the wallet allocated in $RCAT and will deduct 1 (one) $RCAT every 10 minutes to keep it running.
{% endhint %}


# $RCAT Tokenomics

This first agent token is an experiment and a call to gather a smart community around our vision. We will use it to implement token mechanisms that benefit other users' agents.&#x20;

We're planning a comprehensive **Tokenomics** structure for the platform once its Alpha version is live, and early contributors to the **Replicat-One** token will be in our hearts and minds.&#x20;

Incentives matter, and we believe in trust built with long-term alignment.

### **Public**

<table><thead><tr><th width="136">Area</th><th width="116">Allocation</th><th width="141">Vesting</th><th>Remarks</th></tr></thead><tbody><tr><td>Uniswap v2 LP Pool</td><td>12.5%</td><td>Unlocked</td><td>Auto-migrated after agent red-pill and paired with 41,600 VIRTUALs.<br><br><strong>Important</strong>: LP is forever locked</td></tr><tr><td>Airdrops</td><td>10%</td><td>Unlocked</td><td><ul><li>Airdropped to vetted Virtual holders with retention up to 440% to avoid supply into the hands of snipers<br></li><li>Grants and community initiatives.</li></ul></td></tr></tbody></table>

### Private

<table><thead><tr><th width="137">Area</th><th width="114">Allocation</th><th width="145">Vesting</th><th>Remarks</th></tr></thead><tbody><tr><td>Team</td><td>10%</td><td>6mo lock, 18mo linear vest</td><td>Team alignment and incentives</td></tr><tr><td>Strategic Round </td><td>10%</td><td>Unlocked</td><td>Sold by the team at the top of the bonding curve valuation of 1.6M USD.<br><br><strong>Important: no tokens were or will ever be sold by the team below a 1.6M USD valuation.</strong></td></tr><tr><td>Advisory</td><td>5%</td><td>6mo lock, 18mo linear vest</td><td>For curent and future advisories for Replicats </td></tr><tr><td>Treasury</td><td>32.5%</td><td><p>Locked.</p><p>12mo vest, 3mo cliff</p></td><td>For development, operational costs, and hiring</td></tr><tr><td>Marketing</td><td>15%</td><td><p>Locked.</p><p>12mo vest, 3mo cliff</p></td><td>Related to marketing and community activities aiming for project growth</td></tr><tr><td>Listings</td><td>5%</td><td><p>Locked.</p><p>12mo vest, 3mo cliff</p></td><td>CEX listings and Market Makers</td></tr></tbody></table>

$RCAT was originally deployed on Base alongside $VIRTUAL and later made available on Solana via [deBridge](https://debridge.finance/). We bridged some tokens so they would be available to all users.

**Total supply: 1,000,000,000 $RCAT**

**Base**  - 0x6AF73D4579c70A24D52e4F4b43EeCB2A75019F94

The token unit price should be exactly the same on both chains in USD terms.\
To verify the supply and market cap, please check [BaseScan](https://basescan.org/token/0x6af73d4579c70a24d52e4f4b43eecb2a75019f94).

{% hint style="info" %}
$RCAT is a utility token for our first Agent, Replicat-ONE.
{% endhint %}


# Institutional-Grade Risk Management

### Systematic Execution Over Emotional Trading

Most crypto traders lose money not because they lack information, but because emotions override logic. You know you should buy the dip, but fear holds you back. You know you should take profits, but greed keeps you holding. You know you should rebalance, but the market moves faster than you can act.

Replicat-ONE removes emotion from execution. The agent doesn't feel FOMO. It doesn't panic sell. It executes the same systematic process regardless of whether Bitcoin just pumped 20% or dropped 30%. This consistency is what separates systematic strategies from discretionary trading.

### Forecasting, Not Reacting

Trading bots chase momentum signals and react to price movements. Replicat-ONE forecasts market conditions using quantitative models, then positions portfolios accordingly. The agent analyzes multiple data streams, identifies structural shifts in volatility regimes, and adjusts exposure before major moves occur.

Instead of predicting, Replicats uses probabilistic positioning. The agent doesn't claim to know what Bitcoin will do tomorrow. Instead, it calculates the probability distribution of outcomes, sizes positions based on those probabilities, and manages risk accordingly.

### Factor-Based Selection

Replicat-ONE doesn't pick assets randomly or chase narratives. It uses factor-based selection to identify assets with favorable risk-adjusted characteristics. Momentum, volatility, correlation structure, and on-chain metrics - the agent synthesizes multiple factors to construct portfolios systematically.

When market conditions change, the agent rebalances in response to updated factor signals. This approach captures market opportunities while maintaining disciplined risk management.

Most crypto trading bots execute simple strategies: arbitrage, grid trading, and DCA. They're tools that do what you tell them. Replicat-ONE actively manages your portfolio using institutional frameworks designed for professional capital allocation.

The difference is sophistication. Trading bots follow rules. Smart Wealth Companions implement adaptive strategies that respond to changing market conditions while maintaining consistent risk controls.

<br>


# Building Wealth, Not Chasing Trades

Crypto culture celebrates the big score: the 10x, the perfectly timed entry, the meme coin that went parabolic. But sustainable wealth isn't built on occasional wins; it's built on consistent execution that compounds over time.

The math is simple. A strategy that returns 2% monthly with controlled drawdowns outperforms a strategy that swings between 50% gains and 40% losses. Compounding requires capital preservation. You can't compound what you've lost.

**Replicat-ONE** exists for wealth building, not speculation. The agent prioritizes keeping what you have while capturing systematic growth opportunities. When markets present favorable risk-adjusted setups, it scales in. When risk exceeds acceptable thresholds, it scales out.&#x20;

This disciplined approach generates returns that compound rather than evaporate.

### **The Transition from Trading to Wealth Building**

Most crypto participants start as traders. Chart-watching, position-sizing on conviction, timing entries and exits. This works until it doesn't. Volatility eventually punishes even skilled discretionary traders who can't maintain 24/7 vigilance.

**Smart Wealth Companions** help you transition from active trading to systematic wealth-building. You're not abandoning crypto or its opportunities. You're implementing professional portfolio management that executes continuously while you focus on everything else in your life.

Institutional capital thrives on systematic frameworks, disciplined execution, and stringent risk controls designed to safeguard assets across all market cycles. Replicats delivers this sophisticated infrastructure directly to individual investors committed to serious wealth building.<br>


# Not Another Trading Bot

Most crypto trading bots execute simple strategies: arbitrage, grid trading, and DCA. They're tools that do what you tell them. **Replicat-ONE** actively manages your portfolio using institutional frameworks designed for professional capital allocation.

The difference is sophistication. Trading bots follow rules. Smart Wealth Companions implement adaptive strategies that respond to changing market conditions while maintaining consistent risk controls.

Individual investors can now access the same institutional-grade infrastructure that institutional funds rely on, available via a monthly subscription.


# What's Next: New Agents and Strategies

Replicat-ONE is our first agent, focused on crypto assets. We're building additional agents that will provide access to tokenized stocks, the S\&P 500, gold, and other real-world assets (RWAs). On-chain rails enable global access to these assets regardless of geographic restrictions.

Each new agent will operate on the same quantitative foundation, but will be optimized for its specific asset class. Same custody model. Same transparency. Broader opportunities.


# FAQ


# Important Notice

{% hint style="info" %}
Important Notice to Replicats.AI Users
{% endhint %}

Important Notice to Replicats Users

Thank you for exploring the Replicats platform! Please note that this project is currently in its developmental stage. Features are being deployed on an experimental basis to encourage community engagement and gather valuable user feedback.

We want to emphasize that several aspects of the platform, including technical features, the business model, terms and conditions, and other relevant items, are subject to updates and additions. These updates will occur over the coming days and weeks as we work to refine and enhance the platform.

Your input is essential to the success of Replicats. We encourage you to share your feedback, suggestions, and ideas to help us create the best possible experience for everyone.<br>

Thank you for your understanding and support during this exciting phase of development.<br>

The Replicats Team


# Legal

### **Disclaimers**

**Replicats** is an AI platform that provides smart wealth companions for anyone transitioning to sophisticated wealth building. Our flagship agent, **Replicat-One**, has been actively managing portfolios since February 2025, utilizing institutional-grade optimization techniques and on-chain execution.

\
Please note that **Replicat-One** and **Replicats** are currently in development and testing. Features are being deployed experimentally to encourage community engagement and gather user feedback. We emphasize that several platform aspects, including technical features, the business model, terms and conditions, and other relevant items, are subject to change.

\
**Replicats** and **Replicat-ONE** are personalized for each user based on the agent's baseline strategy.


# Agent Building

Guide to creating sophisticated trading agents using Replicats' intuitive tools.

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FL8CA9VpOqbaUoU8LtRco%2FSOON%20(4).png?alt=media&amp;token=4fae7ce3-0124-4652-bd68-a82449a6a49b" alt=""><figcaption></figcaption></figure>


# Agent Management

Tools and best practices for managing your autonomous trading agents.

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FL8CA9VpOqbaUoU8LtRco%2FSOON%20(4).png?alt=media&amp;token=4fae7ce3-0124-4652-bd68-a82449a6a49b" alt=""><figcaption></figcaption></figure>


# First Agent: Replicat-ONE

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FZ5fJoN6TX2I869XXn2EP%2FScreenshot%202025-05-26%20at%2010.35.36.png?alt=media&amp;token=2725ee86-5218-48c8-9a57-c1aa9d9b31b5" alt=""><figcaption></figcaption></figure>

**Replicat-ONE** is the first Agent on the Replicats platform.

This sophisticated hunter combines technical and fundamental analysis to track and capture opportunities and market movement while its built-in risk management keeps your assets safe.&#x20;

Users can clone and deploy Replicat-ONEs on their wallets, allowing them to execute their crafted strategies with pinpoint precision.

### Core Strategy

**Replicat-ONE** is being built to outperform BTC in the long run.

***

**Replicat-ONE** is currently live and trading on [**Base** following a list of pre-determined assets](https://paragraph.com/@replicatsai/replicat-one-trading-strategy).

You can run Replicat-ONE's backtest strategy locally with this file on our [GitHub](https://github.com/replicatsai/replicats-sample-backtest) repository.

{% hint style="info" %}
**Replicat-ONE** token launch used the Virtuals launchpad with the RCAT/VIRTUAL pair.\
We added several LPs with USDC on Base and Solana.
{% endhint %}


# Contract Addresses

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FrV4PrJNiZNFFrZgceaDu%2FXpost-BaseSolana.png?alt=media&amp;token=b4e80624-be31-4e66-826e-97d0a581f0ed" alt="" width="375"><figcaption><p>$RCAT live on Base and Solana</p></figcaption></figure>

$RCAT was originally deployed on Base paired with $VIRTUAL and was later made available on Solana through [deBridge](https://debridge.finance/). We bridged some tokens so they would be available to all users.

**Total supply: 1,000,000,000 $RCAT**

**Base**  - 0x6AF73D4579c70A24D52e4F4b43EeCB2A75019F94

**Solana** - 5qaZLZ6vHL4mXLz7sZmWnXjhkzX1tPZLGLkgGc6PJbGX<br>

The token unit price should be exactly the same on both chains in USD terms.\
To verify the supply and market cap, please check [BaseScan](https://basescan.org/token/0x6af73d4579c70a24d52e4f4b43eecb2a75019f94).


# Why Replicats

Personal Investment Experts In Your Pocket.

Building the next-gen Portfolio Management tool


# Current State of Crypto Trading

An analysis of current market challenges and limitations of existing trading approaches in the cryptocurrency space.

Crypto never sleeps. Currencies and tokens trade around the clock, often experiencing **explosive** volatility in a matter of hours—or even minutes.

* **Continuous Flux**: Traders must constantly monitor multiple exchanges, DeFi protocols, and liquidity pools.
* **Heightened Complexity**: On-chain data, social sentiment, and macro news can instantly shift market sentiment.

### The Reality of “AI” in Crypto Today

Many platforms advertise as “AI-driven,” but most rely on **basic** sentiment analysis or scraping tweets to determine market direction.

* **Over-Reliance on Language Models**: LLMs alone can generate a decent read on social chatter but **struggle** to process **extensive historical** data or the **complex** nature of on-chain relationships.
* **Shallow Predictions**: Without deeper numerical or topological insights, these systems often yield **generic** or **lagging** recommendations.

***

### The Need for a Multi-Modal Approach

The crypto ecosystem spans **time-series** (price and volume), **graph data** (transactions, wallet interactions), and **unstructured** data (tweets, forum posts).

* **Holistic Analysis**: True market intelligence demands **combining** specialized models for each data type.
* **Scalability**: As new tokens and protocols emerge, your AI system must adapt rapidly, ingesting newly minted on-chain data and trending social content.

***

### Where We Stand

Traders and institutions are searching for **robust** tools that move beyond surface-level insights. In this evolving landscape, **autonomous Agent frameworks** are fast becoming the standard for advanced, **round-the-clock** trading strategies.


# Our Approach

Our unique perspective on combining representation learning with autonomous agents for superior trading outcomes.

## Core Thesis

The cryptocurrency market's complexity demands more than simple automation or language model analysis. At Replicats, effective trading requires a deep structural understanding of market dynamics, achieved through sophisticated representation learning and autonomous execution.

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2F9vkUYazo7uEsXAOAIIE0%2FRCAT_Framework2.png?alt=media&amp;token=58fa9635-5263-46d4-b2d1-8ab8ee09774b" alt=""><figcaption><p>Replicats Platform</p></figcaption></figure>

#### The Representation Learning Advantage

While others focus on surface-level patterns or simple metrics, we look deeper into market structure. Through representation learning, we discover and encode the underlying patterns and relationships that drive market behavior:

$$
\mathcal{L}(\theta) = \mathbb{E}*{x\sim p(x)}\[-\log p*\theta(x)] + \beta \text{KL}(q\_\phi(z|x)||p(z))
$$

This mathematical foundation allows us to capture complex market dynamics in a way that's both rigorous and practically effective. Rather than treating market data as simple time series or disconnected events, we model the deep structural relationships that drive market behavior.

***

### Beyond Buzzwords: Real AI Expertise

In a market saturated with AI buzzwords, Replicats stands apart through deep expertise in representation learning. Our approach isn't about jumping on the latest AI trend but applying proven mathematical foundations to solve real trading challenges.

#### The Foundation of Our Intelligence

At the core of our platform lie two specialized foundation models:

1. **Time Series Foundation Model** Our temporal model captures market dynamics through sophisticated time series analysis.
2. **Graph-based Foundation Model** Our heterogeneous graph transformer architecture models complex market relationships through a knowledge graph containing millions of vertices, capturing: token interactions, market participant behaviors, on-chain dynamics, and wallet patterns.

> "Market relationships aren't just about price correlations—they're about understanding the complex web of interactions between protocols, tokens, and market participants."

#### Model Fusion

Our most significant innovation lies in how we combine these models. Through a sophisticated fusion architecture:

$$
z = \text{Fusion}(h\_\text{time}, h\_\text{graph}) = \text{Encoder}(\text{Concat}(h\_\text{time}, h\_\text{graph}))
$$

This allows us to capture temporal dynamics and structural relationships in a unified representation, providing a complete understanding of market conditions.

### The Power of DAG-Based Workflows

Rather than relying on black-box solutions, Replicats employs a sophisticated Directed Acyclic Graph (DAG) workflow engine. This approach provides:

#### Modular Intelligence

Each node in the DAG can represent:

* Data transformation steps
* Model predictions
* Trading decisions
* Risk assessments

#### Conditional Execution

The DAG enables sophisticated trading logic:

```
IF (prediction_confidence > threshold) AND (risk_assessment = acceptable):
    THEN execute_trade()
    ELSE reassess_position()
```

This structure allows for complex, multi-step strategies while maintaining clear logic and accountability.

### Strategic Use of Language Models

While many platforms try to solve everything with LLMs, we take a more nuanced approach. LLMs serve a specific role in our architecture:

#### Where LLMs Excel

* Interpreting user intentions
* Summarizing market narratives
* Providing strategy explanations

#### Where Specialized Models Take Over

* Price prediction
* Pattern recognition
* Risk assessment
* Trade execution

This hybrid approach allows us to leverage the best of both worlds—natural language understanding where it matters, and specialized mathematical models where precision is crucial.

### The Power of Deep Understanding

Our representation learning approach enables:

#### Complex Pattern Recognition

* Early detection of market regime changes
* Understanding of cross-token influences
* Recognition of emerging market structures

#### Predictive Intelligence

Through comprehensive market understanding:

* Pattern recognition across multiple timeframes
* Complex relationship identification
* Market regime classification and prediction

#### Risk Management

Sophisticated risk controls through:

* Multi-dimensional risk assessment
* Dynamic position sizing based on market structure
* Cross-token correlation analysis

### Looking Forward

The crypto market's complexity continues to grow, but so does our capability to understand it. Through our commitment to deep technical expertise and practical trading solutions, Replicats remains at the forefront of autonomous trading technology.

> "The future of trading isn't about replacing human intelligence—it's about extending it through sophisticated representation learning and autonomous execution."


# Trading-Specific Agents

Exploring the paradigm shift from manual trading to autonomous agents and why this transformation is crucial for modern crypto trading.

### The Evolution of Trading Intelligence

The cryptocurrency market presents a unique challenge in financial history: a 24/7 global market with unprecedented data complexity and execution speed requirements. This environment has pushed beyond the capabilities of both human traders and traditional algorithmic systems, creating the perfect conditions for the rise of AI agents.

But what exactly constitutes an AI Agent in cryptocurrency trading? Far more than simple automation, a trading Agent represents an autonomous system capable of perceiving market conditions, making complex decisions, and executing actions independently – all while maintaining alignment with its defined strategy and risk parameters. This autonomy and sophistication represent a fundamental shift from traditional trading approaches.

### Why Traditional Approaches Fall Short

The limitations of traditional trading approaches become particularly apparent in cryptocurrency markets. Despite their intuition and experience, human traders cannot process the vast amounts of data generated across multiple chains and protocols. Even the most dedicated trader cannot maintain consistent 24/7 market coverage or process thousands of market signals simultaneously.

While capable of continuous operation, traditional algorithmic trading systems typically rely on fixed rules and parameters. They lack the adaptability required for crypto markets, where market conditions can shift dramatically based on complex technological, social, and economic interactions.

### The Agent Advantage

The true power of agents lies in their ability to combine multiple forms of intelligence. Modern trading agents can simultaneously:

* Process multiple data streams in real-time, from price action to on-chain metrics&#x20;
* Understand complex market relationships through sophisticated modeling
* Execute precise trading strategies without emotional bias&#x20;
* Adapt to changing market conditions autonomously

Perhaps most importantly, agents can maintain consistency in their operations while still being flexible in their strategy. This seemingly paradoxical capability comes from their foundational architecture—they follow consistent decision-making processes while allowing the inputs and parameters of those processes to evolve with market conditions.

***

### Beyond Simple Automation

What truly sets modern agents apart is their ability to understand market context. Unlike simple automated systems that follow predefined rules, sophisticated agents can:

Understanding Market Context: Modern agents don't just react to predefined triggers; they understand the broader market context through multiple specialized models. This might involve analyzing social sentiment alongside technical indicators, or combining on-chain metrics with traditional market data to form a comprehensive view of market conditions.

Dynamic Strategy Adaptation: Rather than following fixed strategies, agents can adjust their approach based on changing market conditions. This adaptation isn't random – it's guided by sophisticated models that understand market regimes and can adjust parameters accordingly.

Risk Management 2.0: Traditional risk management often relies on simple stop-losses or position sizing rules. Agent-based risk management can be far more sophisticated, considering market liquidity, correlation risks, and systemic market risks when making decisions.

***

### The Future of Trading

The movement toward agent-based trading isn't just a technological trend – it's a necessary evolution driven by market complexity. As crypto markets become more sophisticated, with increasing interconnections between protocols and growing data complexity, the role of intelligent agents will become increasingly central to successful trading strategies.

Consider the challenge of trading in a market where value flows between protocols, chains, and layers. A human trader or simple algorithm might struggle to track these complex interactions, but an Agent can monitor these relationships continuously, identifying opportunities that arise from these interconnections.

This isn't to say that agents will completely replace human traders. Instead, they represent a new paradigm where traders can encode their strategies and market understanding into autonomous systems that can execute consistently and adapt intelligently. The future of trading lies not in choosing between human and machine, but in creating sophisticated agents that can extend human trading capabilities far beyond their natural limitations.


# Beyond LLMs

Why representation learning surpasses pure LLM approaches in cryptocurrency trading.

### The Limits of Pure Language Models

The recent surge in Large Language Model applications has led many to view them as a universal solution for complex problems. In the cryptocurrency trading space, numerous platforms have emerged claiming to leverage LLMs for market analysis and trading decisions. However, this approach fundamentally misunderstands both the capabilities of LLMs and the nature of financial markets.

LLMs excel at pattern recognition in natural language and can engage in sophisticated reasoning about qualitative information. However, they face significant limitations when dealing with the quantitative, real-time nature of financial markets. These limitations stem from their fundamental architecture and training approach.

Consider the basic architecture of a transformer-based LLM:

$$
\text{Attention}(Q, K, V) = \text{softmax}\left(\frac{QK^T}{\sqrt{d\_k}}\right)
$$

While this mechanism is powerful for natural language processing, it presents several critical limitations for market analysis:

#### Context Window Constraints

Even with recent advances in context window sizes (from 2048 tokens in early models to 32k or more in recent ones), LLMs still cannot maintain the comprehensive market history needed for sophisticated trading decisions. A single day of high-frequency market data can easily exceed these limits.

#### Numerical Precision

LLMs process numbers as tokens, leading to potential precision loss. Consider a simple price series:

$$
P = {1234.5678, 1234.5679, 1234.5680}
$$

To an LLM, these numbers are just tokens, making it difficult to perform precise calculations or recognize subtle patterns that might be crucial for trading decisions.

#### Real-time Processing Limitations

The computational overhead of processing large prompts through an LLM creates significant latency:

$$
\text{Total Latency} = t\_{\text{prompt}} + t\_{\text{processing}} + t\_{\text{generation}}
$$

In fast-moving markets, this latency can mean the difference between a profitable trade and a missed opportunity.

### Market Complexity and Data Structure

Financial markets, particularly in cryptocurrency, generate data that is inherently structured and relational. This data includes:

1. Time series data (prices, volumes, metrics)
2. Graph data (transaction networks, token relationships)
3. Event data (smart contract interactions, protocol updates)

Each of these data types requires specialized processing approaches that align with their mathematical structure:

$$
\text{Time Series}: X\_t = f(X\_{t-1}, X\_{t-2}, ..., X\_{t-n}) \\
\text{Graph Structure}: G = (V, E, A) \\
\text{Event Sequence}: E = {(e\_1, t\_1), (e\_2, t\_2), ..., (e\_n, t\_n)}
$$

LLMs, designed for natural language, lack the specialized architectures needed to process these data structures efficiently and accurately.

### The Role of Representation Learning

Instead of forcing all market data through the bottleneck of language models, a more effective approach is to use specialized models that can learn appropriate representations for each type of market data. This allows for:

1. Preservation of data structure and relationships
2. Efficient processing of numerical information
3. Capture of temporal dependencies
4. Understanding of market microstructure

The key is to let each type of data be processed by architectures designed for its specific characteristics:

$$
h\_\text{time} = f\_\text{temporal}(X\_t) \\
h\_\text{graph} = f\_\text{graph}(G) \\
h\_\text{event} = f\_\text{event}(E)
$$

These specialized representations can then be combined through sophisticated fusion techniques while maintaining their essential properties.

### The Future of Market Analysis

The future of market analysis lies not in forcing all data through language models, but in developing specialized architectures that can capture the true complexity of market behavior. This requires a deep understanding of both the mathematical structures underlying different types of market data and the computational architectures best suited to processing them.

LLMs still have a role to play, particularly in:

* Interpreting market news and sentiment
* Providing human-friendly explanations of market behavior
* Processing qualitative market information

However, they should be seen as one tool in a broader arsenal, not as a universal solution to all market analysis challenges.


# The Limits of Pure Language Models

Understanding why traditional LLM-based approaches fall short in complex market analysis.

### The Tokenization Problem

The fundamental issue with using LLMs for market analysis begins at the tokenization level. When processing numerical data, LLMs break numbers into tokens based on their characters rather than their mathematical significance. Consider a simple price sequence:

$$
P = {19857.32, 19857.33, 19857.34}
$$

To an LLM, this might be tokenized as:

```
['19', '857', '.', '32'], ['19', '857', '.', '33'], ['19', '857', '.', '34']
```

This tokenization destroys the numerical relationships that are crucial for market analysis. The model has no inherent understanding that these represent a monotonically increasing sequence with constant differences. Instead, it must try to reconstruct this understanding through pattern matching across tokens.

### Computational Inefficiency

The attention mechanism in transformer-based LLMs, while powerful for natural language, becomes computationally inefficient for numerical analysis:

$$
\text{Complexity} = O(n^2 d)
$$

Where n is the sequence length and d is the embedding dimension. For high-frequency market data, this quadratic complexity becomes prohibitive. A single day of minute-level data for multiple market indicators can easily exceed practical processing limits.

### The Hidden State Problem

LLMs lack explicit state management for tracking market conditions. Their understanding of state must be encoded in the attention patterns:

$$
\text{Attention}(Q\_t, K\_{1:t}, V\_{1:t})ht​
$$

This makes it difficult to maintain consistent tracking of:

* Position sizes
* Portfolio values
* Running statistics
* Risk metrics

### Temporal Understanding Limitations

Market data has explicit temporal structure that LLMs struggle to capture:

$$
\text{Auto-correlation}: R(\tau) = \mathbb{E}\[(X\_t - \mu)(X\_{t+\tau} - \mu)] \\
\text{Volatility clustering}: \sigma\_t^2 = \alpha\_0 + \alpha\_1 r\_{t-1}^2 + \beta\_1 \sigma\_{t-1}^2
$$

These temporal dependencies require specialized architectures that can:

1. Maintain explicit time awareness
2. Process multiple timeframes simultaneously
3. Capture regime changes
4. Model temporal dependencies directly

### Context and Causality

LLMs process market data as a sequence of tokens without understanding causality:

$$
p(x\_t|x\_{1:t-1}) \neq p(x\_t|\text{Relevant}(x\_{1:t-1}))
$$

This leads to:

* Spurious correlations
* Inability to distinguish cause from effect
* Poor handling of regime changes
* Limited understanding of market microstructure

### Real-world Impact

These limitations manifest in practical trading scenarios:

1. **Delayed Reactions**: The processing overhead leads to missed opportunities
2. **Inconsistent Analysis**: The same market condition can yield different interpretations
3. **Poor Risk Management**: Inability to maintain consistent risk metrics
4. **Resource Inefficiency**: High computational cost for basic market analysis

The solution isn't to abandon LLMs entirely, but to recognize their appropriate role within a broader market analysis framework. They excel at:

* Processing market news
* Sentiment analysis
* Strategy description
* Explaining complex market events

But they should not be the primary engine for:

* Price prediction
* Risk calculation
* Portfolio optimization
* Trade execution


# Why Representation Learning Matters

The mathematical and practical advantages of representation learning in financial markets.

### The Power of Learned Representations

Representation learning addresses a fundamental challenge in market analysis: how to transform raw market data into meaningful, actionable features. Unlike predetermined features or LLM embeddings, learned representations capture the inherent structure of market data:

$$
\phi: \mathcal{X} \to \mathcal{H}
$$

Where $$\mathcal{X}$$ is the space of raw market data and $$\mathcal{H}$$ is a learned representation space that captures meaningful market dynamics.

### Mathematical Foundations

The power of representation learning comes from its ability to capture complex market structures. Consider a market with multiple assets and various types of relationships. We can model this as a heterogeneous graph:

$$
\mathcal{G} = (\mathcal{V}, \mathcal{E}, \mathcal{A}, \mathcal{R})
$$

Where:

* $$\mathcal{V}$$ represents vertices (assets, traders, protocols)
* $$\mathcal{E}$$ represents edges (relationships)
* $$\mathcal{A}$$ represents vertex attributes
* $$\mathcal{R}$$ represents relationship types

Through representation learning, we can learn embeddings that preserve the essential structure of this market graph:

$$
h\_v = f\_\phi(v, \mathcal{N}(v)) \\
\text{where }\mathcal{N}(v) = \text{ neighborhood of vertex }v
$$

### Temporal Dynamics

Market behavior is inherently temporal. Representation learning allows us to capture these dynamics through specialized architectures:

$$
h\_t = f\_\phi(x\_t, h\_{t-1}) \\
\text{where }h\_t\text{ captures market state at time }t
$$

This allows for:

1. Multi-scale temporal patterns
2. Regime detection
3. Trend analysis
4. Volatility modeling

### The Information Bottleneck

Representation learning operates on the principle of the information bottleneck:

$$
\min\_{p(h|x)} I(X; H) - \beta I(H; Y)
$$

Where:

* $$I(X;H)$$ is the mutual information between input and representation
* $$I(H;Y)$$ is the mutual information between representation and target
* $$β$$ controls the trade-off between compression and prediction

This framework ensures that learned representations:

* Capture relevant market information
* Discard noise
* Maintain predictive power
* Generalize well to new conditions


# Replicats' Hybrid Approach

How we combine specialized models with targeted LLM usage for optimal trading decisions.

### The Power of Specialization

At Replicats, we recognize that no single technology can effectively capture all aspects of cryptocurrency markets. Instead, we've developed a hybrid approach that combines specialized models, each designed for specific aspects of market analysis, with selective use of LLMs where they add value.

### Foundation Models

Our approach centers on two specialized foundation models that form the core of our market understanding:

#### Graph Model

Our graph transformer processes market relationships through a sophisticated knowledge graph. This model captures:

* Token interactions and dependencies
* Protocol relationships and integrations
* Market participant behaviors

The graph model enables us to understand market structure beyond simple price relationships, providing insights into the complex web of interactions that drive cryptocurrency markets.

### Intelligent Integration

What sets our approach apart is how we combine these specialized models with other technologies:

#### Model Fusion

We combine insights from different models through a sophisticated fusion architecture. This allows us to:

1. Merge temporal and structural insights
2. Maintain the strengths of each model
3. Create unified market representations
4. Enable consistent decision-making

#### Strategic Use of LLMs

Unlike platforms that try to solve everything with LLMs, we use them selectively where they add real value:

**Where We Use LLMs:**

* Natural language interface for strategy definition
* Market narrative analysis and summarization
* Strategy explanation and reasoning
* Processing qualitative market information

**Where We Use Specialized Models:**

* Price prediction and pattern recognition
* Risk assessment and portfolio optimization
* Market structure analysis
* Real-time decision making

### Looking Forward

The future of cryptocurrency trading lies not in relying on a single technology but in intelligently combining specialized approaches. Our hybrid architecture provides a foundation for continuous innovation while maintaining the reliability and performance needed for effective trading.

Through this approach, we deliver a trading platform that combines the best of multiple technologies - the precision of specialized models, the flexibility of LLMs, and the reliability of traditional trading systems - all working together to create a more effective trading solution.


# Platform Architecture

A comprehensive overview of Replicats’ core technological components and how they work together.

Replicats is built on three core pillars:

1. An **Agent Framework** that organizes AI-driven tasks via a flexible DAG-based workflow.
2. A **Wallet System** that ensures secure, autonomous asset management for each Agent.
3. A **Trading Engine** that executes trades precisely, handling market signals in real time.

These components function as a **cohesive whole**, enabling Replicats to deliver **intelligent, autonomous** trading solutions that seamlessly blend advanced AI with robust blockchain integration.


# Agent Framework

The Agent Framework represents the intelligence layer of Replicats, orchestrating complex trading strategies through a sophisticated DAG-based workflow engine.

<figure><img src="https://235478734-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM68nyZXNqbdb4AcPAEYf%2Fuploads%2FI0WNa3NHtOCZjiAIMU9X%2FRCAT_Framework%20(1).png?alt=media&amp;token=997553ac-9cf5-4009-a4b7-5437b9ba5ec2" alt=""><figcaption><p>Replicats Inteligence Framework (RIF)</p></figcaption></figure>

### DAG-Based Orchestration

* **Directed Acyclic Graph (DAG) Foundation**\
  Our Agent framework uses a DAG to coordinate tasks such as data ingestion, model inference, risk checks, and trade execution.
  * Each **node** in the DAG represents a specialized step (e.g., reading time series data, querying the knowledge graph, or calling a predictive model).
  * The **edges** define how data flows from one step to another without circular dependencies.
* **Conditional & Event-Driven Logic**
  * Agents can branch into different paths based on **probabilistic triggers** (e.g., a predicted 80% chance of price rally) or deterministic thresholds (e.g., price surpasses $30,000).
  * This flexibility allows for **complex strategies**, from simple rebalancing rules to multi-step yield farming sequences.

### Multi-Model Integration

* **Specialized AI Models for Triggers**
* **Targeted LLM Usage**
  * While specialized models handle the bulk of predictions, we still leverage LLMs (via DSPy) for **natural language queries**, high-level summaries, or user interactions.
  * The DAG calls the LLM steps only where **human-like reasoning** adds value, keeping **cost and latency** under control.


# Wallet System

Understanding our secure, autonomous wallet architecture and customizable access policies.

Replicats provides non-custodial wallets not only for end users but also for Agents, as:

### **1- Dynamic for users**

**Replicats** uses **Dynamic** to offer a state-of-the-art custodial and key-management solution, guaranteeing all users a secure and non-custodial wallet.

* Private keys are **never** exposed to users or third parties
* Dynamic handles encryption and signing in a **protected environment**.
* For more info, please refer to [Dynamic's documentation](https://docs.dynamic.xyz/introduction/welcome).

### **2 - Crossmint for Agents**

Each Agent built with Replicats has a trading wallet attached to trade 100% autonomously.

* Crossmint provides wallets.
* For reference, please check  [Crossmint's documentation](https://docs.crossmint.com/wallets/quickstarts/agent-wallets).

***

### Autonomous Yet Configurable

* **Custom Wallet Assignments**
  * Users can define **each Agent's wallet**, effectively segregating funds or strategies.
  * For example, a **high-risk** DeFi farming Agent might have a wallet separate from a **conservative** spot trading Agent.<br>
* **Access Policy Options**
  * Full autonomy: The Agent can sign and send transactions automatically.
  * Partial autonomy: The Agent requests user approval (e.g., SMS, Telegram) before sending large or risky trades.
  * **Multi-Chain Support**: Agents can manage EVM-compatible addresses and wallets on blockchains like Solana.


# Trading Engine

How Replicats processes market signals and executes trades with precision and reliability.

The Trading Engine represents the execution layer of Replicats, handling the actual interaction with markets and ensuring efficient trade execution.

### Execution Framework

#### Order Management

The engine handles:

* Order creation and validation
* Execution strategy selection
* Transaction monitoring
* Settlement confirmation

#### Market Integration

Supports multiple execution venues:

* DEX integration
* AMM interaction
* Cross-protocol execution
* Smart order routing

{% hint style="info" %}
Work in progress
{% endhint %}


# Data Infrastructure

The backbone of our platform: how we collect, process, and analyze market data.

Replicats' data infrastructure combines multiple streams of market data to create a comprehensive view of the cryptocurrency ecosystem. Through sophisticated processing pipelines and a robust knowledge graph architecture, we transform raw data into actionable market intelligence.

### Knowledge Graph

Our knowledge graph serves as the foundation of our market understanding, modeling the complex relationships between tokens, protocols, and market participants. This graph structure continuously evolves as new relationships and patterns emerge in the market, capturing both explicit connections and implicit relationships discovered through our analysis.

### Data Sources

We process three main categories of data: market data, on-chain data, and alternative data. Market data includes traditional metrics like price and volume, while our on-chain analysis captures transaction flows and smart contract interactions. Alternative data encompasses social sentiment, developer activity, and community engagement signals.

Through our integration layer, these diverse data sources are combined into a unified view of the market, enabling sophisticated analysis and pattern recognition. Our real-time processing capabilities ensure that new information is quickly incorporated into our market understanding.

### Partnership Opportunities

We actively seek partnerships with data providers to enhance our market coverage and analytical capabilities. Whether you're providing alternative data, on-chain analytics, or specialized market metrics, we're interested in exploring how we can work together to improve market understanding.

For potential partners, we offer integration with our sophisticated analytics platform and the opportunity to collaborate on developing new market insights. Interested parties are encouraged to reach out to our partnerships team.

### Looking Forward

As markets evolve, so does our infrastructure. We continuously enhance our capabilities through new data sources, improved processing techniques, and advanced analytics. Our goal is to maintain a comprehensive and accurate view of the cryptocurrency market while providing reliable, actionable insights for our users.