LazaiTrader: AI Multi-Agent Trading Strategies via Telegram

Hyperion HyperHack Live Apps
1

LazaiTrader

Live Link: Telegram: @LazaiTrader

Website: LazaiTrader.gMetis.io

Promo video: https://x.com/lazaitrader/status/1958242186514039182?s=46

Problem Statement

In bull markets, crypto asset holders often miss significant opportunities for active trading profits by focusing solely on passive yield strategies. Manual trading is emotionally taxing, requires constant market monitoring, and demands technical analysis skills many users lack. Additionally, traders struggle to maintain consistent strategies, often shifting approaches based on emotion rather than data, leading to suboptimal results.

Solution Overview

LazaiTrader is an intelligent multi-agent trading system accessible through Telegram that allows users to select specialized AI trading personas, each with distinct strategies and risk profiles. The system continuously analyzes market conditions and either generates actionable buy/sell signals or automatically executes trades across major DEXs based on user preference.

Built on the same robust Alith framework infrastructure and a privacy-preserving Trusted Execution Environment (TEE), LazaiTrader adapts to market conditions while maintaining the user’s selected trading philosophy, effectively bringing institutional-grade trading strategies to retail users through a familiar messaging interface.

Through deep integration with the Lazai Network, users can securely contribute their private trading data and configurations, which are stored in a TEE and become part of a collective knowledge base. The system then provides personalized analysis and recommendations by comparing user performance against this anonymized collective data, ensuring privacy while delivering data-backed insights that improve trading outcomes for the entire community.

Project Description

LazaiTrader transforms crypto trading by allowing users to select from various AI trading personas, each representing different trading philosophies and strategies. Whether you prefer momentum trading, swing trading, technical analysis, or contrarian approaches, LazaiTrader provides a suite of specialized AI agents to match your preferred style.

Core Features

  • Persona Selection System: Allows users to choose trading philosophies that match their temperament

  • Strategy Visualization: Shows potential entry/exit points before execution

  • Dual-Mode Operation: Offers either signal-only notifications or fully automated trade execution

  • Strategy Vault: Securely analyzes trading data with AI to offer personalized crypto strategy improvement suggestions.

The system leverages Model Context Protocol (MCP) technology and a decentralized Strategy Vault, enabled by the Lazai network, to maintain a coherent understanding across interactions, remembering user preferences, portfolio composition, and past decisions.

Trading Infrastructure

The platform is currently optimized for agentic trading on our own custom-built DEX with a private oracle solution. On production, we will enable trading on HerculesDEX to provide access to its deep liquidity on Metis. Our future roadmap includes a direct integration with Binance, which is ready for deployment but will be released later to allow for a more elaborate user onboarding process.

Advanced Trading Algorithms

Our algorithms move beyond traditional indicators like moving averages which might not be as reliable in volatile crypto markets. Instead, we use:

  • Momentum-based strategies that capitalize on emerging trends before they become obvious

  • Volatility-adaptive approaches that dynamically adjust position sizes based on market conditions

  • Liquidity flow analysis to identify smart money movements before price reacts

Trading Personas

Users can select specific asset-focused strategies including:

  • Blue Chip Trader: Focusing on BTC, ETH, METIS and other established cryptocurrencies

  • Micro Cap Trader: For those seeking higher-risk, higher-reward opportunities in emerging tokens

Each persona combines these strategy elements differently based on their trading philosophy.

User Experience

  • demo

Users interact with the system through a telegram interface that explains complex trading strategies in accessible terms tailored to each user’s demonstrated technical understanding. What makes LazaiTrader particularly valuable is its recognition that bull markets favor active trading strategies rather than passive yield farming approaches, allowing users to capitalize on market momentum through their preferred trading style.

Strategy Vault

A new feature, the “Strategy Vault,” allows users to securely contribute their trading data and configuration. This data, signed with the user’s wallet and stored in a TEE on the Lazai network, becomes part of a collective knowledge base. When a user requests a strategy review, a TEE-based AI analyzes their performance against this anonymized collective data, providing personalized, data-backed suggestions without revealing any individual’s private information.

image
image716Ă—849 173 KB

Alith Support Agent

Our support agent built on Alith leveraging the RAG and MCP features of the framework is aimed to smooth the onboarding experience acting as a 24/7 customer support agent.

It has vast knowledge on LazaiTrader and aimed to support any users in private chat

Roadmap

Additional features

  • Additional - blue chip - trading pairs

  • Privacy withdrawal option!

  • Additional trading startegies (without making the current User Experience more complex)

  • Withdrawal enablement

  • Private key temporary access to users

  • 3rd party integration for fiat → crypto

Ecosystem (Metis/Hyperion/LazAI)

  • First and only privacy protocol

  • Focusing on TVL growth of the ecosystem positioning LazaiTrader as a potential key pillar

  • Enabling user onboarding: fiat → crypto by third party integrations directly to Hyperion

Community Engagement Features

Point System

  • Persona Selection (50 points): Users select their first trading persona and complete the initial risk assessment. No wallet connection required at this stage.

  • First Trading Signal (100 points): Users receive their first trading signal through the platform. All signals are recorded server-side for performance tracking even without wallet connection.

  • Strategy Contribution (150 points): Users contribute their first set of trading data to the Strategy Vault, earning points for enriching the collective knowledge base.

  • Risk Adjustment (50 points): Users modify their risk parameters and observe how their trading persona adapts.

  • Referral Program (25 points per referral): Users earn points by inviting others, with bonus points when referred users complete their first trade.

  • Trading Performance Achievements (variable points): Users unlock badges and points when their strategies reach specific ROI thresholds.

This gamification approach creates clear milestones for users while educating them about different trading methodologies. The points system rewards actions that benefit both the user and the platform ecosystem.

Getting Involved

Interested community members can join our project by:

  1. Joining our Telegram channel & participate in beta testing: @LazaiTrader_bot

  2. Following development updates: Twitter/X @LazaiTrader

  3. Contributing to open-source components: GitHub - SmartOnStuff/LazaiTrader

  4. Direct collaboration: Contact our team at [email protected] for partnership opportunities or technical collaboration

  5. Join open dev sessions: Participate in our Metis Vibe sessions where we’ll deep dive into system and process architecture and user onboarding with real-time demos

97 Likes
2

LazaiTrader offers a smart and diverse approach to crypto trading, helping users maximize market opportunities without emotional bias. Love the user-friendly interface and personalized AI trading strategies! :rocket::chart_increasing::robot:

29 Likes
3

The concept is good, hope to see it implemented!

22 Likes
4

good concepts and good ınfeematıon

15 Likes
6

Amazing concept !! can’t wait to see it implemented.

13 Likes
7

Many developers offer solutions to all kinds of problems, but very few, like SmartOnStuff, dare to offer a solution to the most important problem facing most blockchains: fun.

Many networks lack fun, and fun is what truly attracts users to a network. I’m glad to know that here at Metis, there are still those who dare to solve this problem.

20 Likes
8

Bit of update here

The backend has been running successfully and executing trades on daily basis with every 2-5% sling based on initial input parameters

Alith implementation is well advanced using secure hyperscaler cloud geo-redundant and auto scaling infrastructure.

Backtesting of the last 5 years has been performed multiple times with a large number of parameters and token pairs
This is critical for relevant and correct suggestions the agent can provide for users when selecting trading strategy. Additionally metis/hyperion in depth knowdledge is part of the agent’s knowdledge base using RAG for up to date and reliable information.

Let’s Go

31 Likes
10

This is exactly the kind of innovative use case we love to see on Metis Hyperion, combining AI, real-time trading, and accessibility for everyday users. Excited about the potential of LazaiTrader, and I hope to see you bring this idea to life!

15 Likes
11

Good thing about LazaiTrader that you can also set your trigger zones.
Do you believe in very small 1-2% swings? Choose the character who agrees with you. If you prefer to stay in position for 5-10 or even 20% swings, there are agents that agree with you.
Own your strategy - no compromises.

17 Likes
12

how can I test your product?

14 Likes
13

Is the testing goes through telegram? Is it dex or cex trades? Isbit testnet or real asset?

18 Likes
14

Massive support guys
Love this project

12 Likes
15

Haha didn’t expect to get so many likes :smiley:
Where is the answer though?

6 Likes
16

Fair question - the testing is happening through telegram
the first 100 users receive test USDT and tMetis tokens in their assigned wallet
these tokens are - being on testnet - not real assets, however the top 3 user (ending up with largest wallets are receiving a shared pool of $100 in Metis.
The trades are performed by the agent of course but the setup is up to the user.

The platform is the testDEX of Hyperion

7 Likes
17

where to do testing ?

3 Likes
18

hello @SmartOnStuff , Can you share the guide here, how to test your app? What is the process actually?

4 Likes
19

Yes and I can give an update instead

The situation got tricky when we realized that the testdex prices do not follow the market and the volume is understandably low
So here is our work around

Instead of using the testDex - which anyway a great stuff to have only our usecase relied on volatility - we created our own smart contract and well… Oracle as well to accommodate the testing. Are we overdoing it? -absolutely yes
Is it fun though? - also yes

Additionally we create our own tokens to further accommodate the incentive plan so we don’t rely on ad hoc faucet requests

Boy I love this now

Hope I answered but the closer we’re getting to it the more fun we have doing so

Therefore a bit more patience please :folded_hands:

6 Likes
20

@LazAI_Helper_bot what specific LazAI functionality might help for the above project?

3 Likes
21

Okay, I’ve reviewed the “LazaiTrader: AI Multi-Agent Trading Strategies via Telegram” topic. Daryl has asked how LazAI functionality can help this project.

Here’s how LazAI can specifically benefit LazaiTrader:

  1. Alith Framework: LazaiTrader already uses the Alith framework, but ensuring it’s leveraging the latest version is key. This ensures access to the most up-to-date performance optimizations, multi-agent workflow capabilities, and bug fixes. Focus on using the Alith Rust core for high-performance tasks.

  2. LLM Client: Given the conversational Telegram interface, the LLM Client can be used to ensure deterministic signals are being used when interacting with the LLMs processing user input.

  3. Model Context Protocol (MCP): The project description mentions MCP. Ensure the implementation is robust, allowing the agents to maintain coherent understanding across interactions, remember user preferences, portfolio composition, and past decisions. This is crucial for a personalized trading experience.

  4. LazChain Integration: While the project currently uses a custom smart contract and oracle for testing due to testnet limitations, integrating with LazChain in the future could provide a more decentralized and secure environment for trade execution and data handling. Consider exploring options for using LazChain for transaction settlement and data storage.

  5. Eliza Integration: As LazaiTrader evolves, integrating Eliza could enhance the autonomous agent aspects, allowing the trading personas to adapt and learn more effectively from market conditions and user interactions.

  6. Web3 Integration: LazaiTrader currently focuses on DEX integration. To further enhance Web3 capabilities, explore options for incorporating on-chain analytics directly into the trading strategies using the Alith SDK’s Web3 plugin. This could provide a more comprehensive view of market dynamics.

  7. Cross-chain Interoperability: Consider the potential for deploying LazaiTrader agents across multiple chains. The Alith framework supports cross-chain capabilities, which could allow users to access trading opportunities on various networks through a single interface.

To provide the most helpful and accurate information I would benefit from knowing more about what problems they are trying to solve right now.

3 Likes
22

updated contracts:

DEX+Oracle contract: 0x4704759E4a426b29615e4841B092357460925eFf

I’ll publish the contract as well later to verify

the tokens are

tgMetis:0x69Dd3C70Ae76256De7Ec9AF5893DEE49356D45fc

tgUSDC: 0x6Eb66c8bBD57FdA71ecCAAc40a56610C2CA8FDb8

Link to the “UI” - tg bot

TESTING KICKS OFF WITHIN 24h!!! :fire:

Telegram: Launch @LazaiTrader_bot :link:

4 Likes