Introduction
MMAGENTS.TECH is a Solana-based platform that lets anyone deploy their own AI trading agent. Each agent runs autonomously on-chain, monitoring markets and executing trades according to its strategy. while you stay in full control of your funds through a non-custodial wallet model. The platform combines off-chain machine learning, real-time on-chain data ingestion, and a verifiable smart contract execution layer to deliver a trading experience that does not require you to watch the charts or write a single line of code.
What these docs cover
These docs are organized into four sections: platform usage (how to deploy and operate agents), core concepts (Power Rank, marketplace, agent tiers), economics (fees, security, risk math), and technical reference (architecture, API, arbitrage mechanics). Every page is self-contained. you can read in order or jump to what you need. Cross-references between pages let you drill down into related topics without losing your place.
If you are new to MMAGENTS, the recommended reading order is Introduction → Getting Started → AI Agents → Power Rank → Fees. This sequence mirrors the path most users follow: understand the platform, deploy your first agent, learn how it earns, then understand the economics. The technical pages (Architecture, API, Arbitrage) can be read in any order and are primarily aimed at developers, integrators, and auditors.
Who these docs are for
The documentation is written to serve four distinct audiences. Each page indicates its primary audience at the top, but every page is readable by every audience. you do not need a developer background to follow the technical pages, and you do not need to be a trader to understand the user-facing pages.
- End users. anyone deploying their first agent through the Telegram bot. You will spend most of your time in Getting Started, AI Agents, Fees, and Security.
- Traders. users comparing agent strategies, sizing capital, planning withdrawals, and evaluating marketplace opportunities. You will benefit from Power Rank, Marketplace, Fees, and the risk-math sections of Security.
- Developers. building on top of the MMAGENTS SDK, integrating smart contracts, or wiring webhooks into a downstream application. You will spend most of your time in Technical Architecture, API & SDK Reference, and Arbitrage & MEV.
- Auditors. reviewing the security model, the trade simulation logic, and the non-custodial architecture. The Security & Risk page and the Technical Architecture page are your primary entry points.
Prerequisites
Before you deploy your first agent, make sure you have the following in place. None of these are difficult to obtain, but having them ready ahead of time will make the deployment process much smoother.
- A Telegram account (the agent lives inside the Telegram bot. there is no separate mobile app or web dashboard to install)
- A funded Solana wallet. at minimum 5 SOL to deploy CIPHER, plus trading capital (5 SOL minimum for CIPHER, more for higher tiers)
- Basic familiarity with crypto transactions (send, receive, confirmations). if you have ever sent SOL or any other crypto between wallets, you are ready
- An exchange account (Coinbase, Binance, Kraken, or any exchange that sells SOL) so you can acquire the SOL needed to deploy and fund your agent
Quickstart
If you just want to get started immediately and learn the details later, follow these five steps. The full walkthrough with screenshots is on the Getting Started page.
- Open Telegram and search
@Mememachine_techbot - Tap Start, then Create Wallet. the bot generates a fresh Solana wallet tied to your Telegram account
- Buy SOL on Coinbase (or any exchange) and send it to the wallet address shown in the bot
- Tap Deploy Agent, choose CIPHER, confirm the 5 SOL fee and wait ~1 minute for on-chain confirmation
- Fund the agent with 5 SOL trading capital, then tap Start Agent. your agent is now live and trading
From wallet creation to a running agent, the entire flow takes about 30 minutes if you already have a funded exchange account. Most of that time is waiting for the SOL transfer from your exchange to confirm on-chain.
How to navigate these docs
Use the sidebar on the left to jump between pages. Each category groups related pages together. On mobile, tap the menu icon in the top-left corner to expand the sidebar. At the bottom of every page you will find links to the previous and next pages in the suggested reading order, so you can read the docs linearly if you prefer.
Throughout the docs you will see several recurring elements: inline code snippets for commands and identifiers, standalone code blocks with syntax highlighting for multi-line examples, equation blocks with a green left border for mathematical formulas, and colored tip boxes for important notes. Tips come in three colors. blue for information, yellow for warnings, and green for best practices.
Where to go next
Depending on your goals, here are the four most useful next pages:
Platform philosophy
MMAGENTS was built on a simple premise: the average person should be able to participate in algorithmic trading without needing to learn how to code, read charts, or sit in front of a screen all day. Traditional algorithmic trading is gatekept by technical complexity. you need to write Python scripts, configure API keys, manage infrastructure, and debug edge cases. MMAGENTS removes every one of those barriers by wrapping the entire experience inside a Telegram bot and letting the AI handle the hard parts.
The platform is intentionally non-custodial. Your private key is encrypted locally on your device and every trade is routed through audited smart contracts. You can stop the agent, withdraw your capital, and walk away at any time, with full control over your funds at every step.
The Power Rank system was designed to make agent performance legible and tradeable. In traditional trading, "track record" is a fuzzy concept. someone claims they made 300% last year, but there is no way to verify it. MMAGENTS solves this by recording every trade on-chain and computing a transparent 0–100 score that directly dictates monthly earnings. This makes agents comparable, tradeable, and valuable as digital assets in their own right.
How MMAGENTS differs from other AI trading platforms
There are several AI trading platforms in the crypto space, but MMAGENTS differs in four key ways. Understanding these differences will help you decide whether MMAGENTS is the right fit for your trading goals.
- Non-custodial by default. Your funds stay in your own wallet. The platform only signs trades you have explicitly authorized through the bot. funds are never deposited into a centralized platform wallet.
- Telegram-native. There is no web dashboard, no mobile app to install, no browser extension. Everything happens inside a Telegram conversation. This makes the platform instantly accessible to anyone who already uses Telegram, with zero onboarding friction.
- Tradeable agents. On other platforms, your trading bot is a sunk cost. you paid for it, you use it, it has no resale value. On MMAGENTS, your agent accumulates a Power Rank that represents its proven track record, and that track record has real market value. You can sell your agent on the P2P marketplace for a multiple of its monthly earnings.
- Trade simulation before execution. Every trade is simulated off-chain before the smart contract executes it. If the simulation shows the trade is unprofitable after gas, slippage, and MEV risk, the trade is skipped. This prevents the agent from bleeding capital on marginal opportunities and is a significant safety advantage over platforms that execute trades blindly.
Key terms you will encounter
Before diving deeper, familiarize yourself with these core terms. A complete glossary is available on the FAQ & Changelog page, but these are the ones you will see most frequently:
| Term | Quick definition |
|---|---|
| Agent | An AI trading bot deployed on Solana that executes trades autonomously |
| PR | Power Rank. a 0–100 score dictating monthly profit (PR × 10 SOL) |
| SOL | Solana. the native cryptocurrency used for all MMAGENTS operations |
| Deploy | The one-time fee to create an agent instance on-chain |
| Fund | Adding trading capital to a deployed agent |
| Non-custodial | You hold your own keys; the platform never controls your funds |
| DEX | Decentralized exchange. e.g. Raydium, Orca, Meteora |
Getting help
If you get stuck at any point, there are three ways to get help. The fastest is the Circle Community, where other users and team members can answer questions in real time. For private account issues (wallet, private key, withdrawal problems), message the admin directly through the Circle community. For developer questions (SDK, API, smart contract integration), the API & SDK Reference page has code examples and error handling guides.
The Solana blockchain advantage
MMAGENTS runs on Solana rather than Ethereum or another Layer 1 for several important technical reasons. Understanding why Solana was chosen helps you appreciate the platform's performance characteristics and limitations. Solana offers dramatically higher throughput than Ethereum (theoretical 65,000 transactions per second vs Ethereum's ~15), significantly lower transaction fees (typically under $0.001 per transaction vs Ethereum's $1–50+ during congestion), and sub-second block confirmation times (400ms vs Ethereum's 12 seconds). These properties are essential for a high-frequency trading platform where every millisecond and every fraction of a cent matters.
Solana's architecture uses a combination of Proof of Stake (PoS) consensus and Proof of History (PoH), a cryptographic clock that enables validators to process transactions in parallel without waiting for global state synchronization. This means that MMAGENTS agents can submit trades and receive confirmation in under a second, which is critical for strategies like QUANTUM's arbitrage where opportunities close within 1–3 seconds. On Ethereum, the same strategy would be impossible because the 12-second block time would mean most arbitrage opportunities are gone before the transaction even confirms.
The trade-off is that Solana has experienced network outages in the past. something Ethereum has not. While Solana's uptime has improved significantly since 2023, the risk of downtime exists and users should be aware of it. During a Solana outage, all MMAGENTS agents pause automatically and resume when the network recovers. Your funds are never at risk during an outage. they remain in the agent's smart contract and are not affected by network downtime. However, the agent cannot trade during the outage, which means missed opportunities.
Understanding risk in algorithmic trading
The platform categorizes and manages several types of risk through its architecture: market risk (managed through position sizing and diversification), liquidity risk (managed through trade simulation and slippage protection), and counterparty risk (minimized through the non-custodial model). The Security & Risk page covers each of these in detail with the mathematical models used to quantify them.
The role of AI in modern trading
The AI models powering MMAGENTS agents represent a significant evolution in retail trading accessibility. Historically, algorithmic trading was the exclusive domain of institutional players. hedge funds, proprietary trading firms, and investment banks with teams of PhD quants, custom-built infrastructure, and direct market access through co-located servers. The minimum capital to participate was in the millions of dollars, and the technical barrier to entry required years of specialized education. MMAGENTS democratizes access to algorithmic trading by packaging institutional-grade AI models into deployable agents that anyone can operate through a Telegram conversation.
The specific AI techniques used include transformer-based neural networks for pattern recognition (similar to the architecture behind GPT models, but trained on market data rather than text), reinforcement learning for trade execution optimization (the agent learns from past trades which execution strategies produce the best outcomes), and Bayesian inference for probability estimation (every trade is evaluated with a confidence score that reflects the model's uncertainty). These techniques are not unique to MMAGENTS. they are the same approaches used by quantitative hedge funds. but MMAGENTS makes them accessible at a fraction of the cost and complexity.
It is important to understand that AI is not magic. AI models are statistical pattern recognizers. they identify correlations in historical data and use those correlations to make probabilistic predictions about future market movements. When the market regime changes in a way that was not represented in the training data, the model's predictions become less accurate. This is why MMAGENTS includes the trade simulation step . it acts as a final safety net that filters out trades where the model's confidence is low, regardless of what the pattern-matching suggests. The combination of AI-driven opportunity detection and deterministic risk filtering is what makes the platform robust across different market conditions.