SendAI Module

The Solana Agent Kit module provides a comprehensive AI-powered toolkit that enables natural language interactions with the Solana blockchain. Users can chat with an AI agent that understands Solana operations and can execute transactions, check balances, and perform complex DeFi operations through conversational interfaces.

Core Functionalities

AI Chat Interface

Advanced chat system with streaming responses, message persistence, and context management

Solana Integration

Direct blockchain interactions through AI commands including transactions, transfers, and DeFi operations

Plugin Ecosystem

Extensible plugin system for adding custom Solana functionality and third-party integrations

Smart Automation

Intelligent automation of complex multi-step Solana operations through natural language commands

Installation & Setup

API Configuration

Configure your OpenAI API key for AI model access:

OPENAI_API_KEY=your_openai_api_key_here

Solana Configuration

Set up Solana RPC and backend endpoints:

HELIUS_STAKED_URL=your_solana_rpc_endpoint
SERVER_URL=your_backend_server_url

Import Module

Import the chat hook and utilities:

import { 
  useChat,
  myProvider,
  generateUUID 
} from '@/modules/solana-agent-kit';

Wallet Integration

Ensure wallet providers are configured:

import { useWallet } from '@/modules/wallet-providers';

AI Model Access: This module requires an OpenAI API key for full functionality. The system gracefully handles missing keys but AI features will be disabled.

Module Architecture

src/modules/solana-agent-kit/
├── hooks/                  # React hooks for chat management
│   └── useChat.ts         # Core chat functionality
├── lib/                   # AI and utility libraries
│   ├── ai/
│   │   └── providers.ts   # AI model configuration
│   └── utils.ts          # Persistence and utility functions
└── index.ts              # Public API exports

Core Hook: useChat

The primary interface for AI-powered Solana interactions:

import { useChat } from '@/modules/solana-agent-kit';
import { useWallet } from '@/modules/wallet-providers';

function AIAssistantChat() {
  const { connected } = useWallet();
  const { 
    messages, 
    input, 
    handleInputChange, 
    handleSubmit, 
    isLoading, 
    error,
    currentOperation,
    append,
    reload,
    stop
  } = useChat({
    id: 'unique-chat-id',
    initialMessages: [],
    onFinish: (message) => {
      console.log('AI response completed:', message);
    },
    onError: (error) => {
      console.error('Chat error:', error);
    }
  });

  if (!connected) {
    return (
      <View style={styles.connectPrompt}>
        <Text>Please connect your wallet to use the AI Agent</Text>
        <ConnectWalletButton />
      </View>
    );
  }

  return (
    <View style={styles.chatContainer}>
      <ChatMessages messages={messages} />
      <ChatInput 
        value={input}
        onChange={handleInputChange}
        onSubmit={handleSubmit}
        loading={isLoading}
        currentOperation={currentOperation}
      />
    </View>
  );
}

Hook Parameters:

id - Unique chat session identifier
initialMessages - Pre-existing message history
onFinish - Callback when AI completes response
onError - Error handling callback

Hook Returns:

messages - Array of chat messages
input - Current input text
handleInputChange - Input change handler
handleSubmit - Message submission function
isLoading - Loading state
error - Error message if any
currentOperation - Current AI operation status
append - Add message programmatically
reload - Regenerate last AI response
stop - Stop current AI generation

AI Model Configuration

The module uses advanced AI models through custom providers:

Chat and Reasoning Models

import { myProvider } from '@/modules/solana-agent-kit';

// Available language models
const models = {
  chat: 'gpt-4o',           // Main chat model
  title: 'gpt-3.5-turbo',  // Title generation
  reasoning: 'gpt-4o',     // Complex reasoning
  fallback: 'gpt-3.5-turbo' // Backup model
};

// Custom provider configuration
const provider = myProvider({
  apiKey: process.env.OPENAI_API_KEY,
  models: models
});

Chat and Reasoning Models

import { myProvider } from '@/modules/solana-agent-kit';

// Available language models
const models = {
  chat: 'gpt-4o',           // Main chat model
  title: 'gpt-3.5-turbo',  // Title generation
  reasoning: 'gpt-4o',     // Complex reasoning
  fallback: 'gpt-3.5-turbo' // Backup model
};

// Custom provider configuration
const provider = myProvider({
  apiKey: process.env.OPENAI_API_KEY,
  models: models
});

Advanced AI Capabilities

Streaming Responses: Real-time message generation
Context Awareness: Maintains conversation history
Reasoning Middleware: Enhanced decision-making
Function Calling: Direct integration with Solana tools
Error Recovery: Automatic retry and fallback mechanisms
Multi-model Support: Different models for different tasks

Solana Agent Capabilities

The AI agent can perform various Solana operations through natural language:

transaction_operations

object

// Example user queries the AI can handle:

"Send 0.5 SOL to 9WzDXwBbmkg8ZTbNMqUxvQRAyrZzDsGYdLVL9zYtAWWM"
"Check my SOL balance"
"What tokens do I own?"
"Swap 1 SOL for USDC"
"Show me my transaction history"

// The AI automatically:
// 1. Parses the user intent
// 2. Constructs appropriate transactions
// 3. Requests user approval
// 4. Executes the operation
// 5. Provides confirmation and details

Available Operations:

Send SOL and SPL tokens
Check balances and portfolios
View transaction history
Sign messages and transactions
Multi-transaction operations

defi_operations

object

// Advanced DeFi queries:

"Create a new token with 1M supply"
"Add liquidity to SOL/USDC pool"
"Stake my SOL for rewards"
"Check my DeFi positions"
"Find the best swap rate for 100 USDC to SOL"

// Integration with other modules:
// - Token creation via Pump.fun or Raydium
// - Liquidity operations via Meteora
// - NFT operations via Tensor
// - Cross-DEX arbitrage opportunities

DeFi Capabilities:

Token creation and management
Liquidity provision and farming
Staking and delegation
Cross-protocol interactions
Yield optimization suggestions

portfolio_management

object

// Portfolio analysis queries:

"Analyze my portfolio performance"
"What's my total portfolio value?"
"Show me my biggest gains and losses"
"Suggest portfolio rebalancing"
"Track my token allocations"

// AI provides:
// - Real-time portfolio valuation
// - Performance analytics
// - Risk assessment
// - Rebalancing suggestions
// - Tax optimization tips

Portfolio Features:

Real-time valuation tracking
Performance analytics
Risk assessment and management
Automated rebalancing suggestions
Tax optimization strategies

Quick Start Examples

import { useChat } from '@/modules/solana-agent-kit';
import { useWallet } from '@/modules/wallet-providers';

function SolanaAIAssistant() {
  const { connected } = useWallet();
  const { 
    messages, 
    input, 
    handleInputChange, 
    handleSubmit, 
    isLoading, 
    currentOperation 
  } = useChat({
    id: 'main-chat',
    onFinish: (message) => {
      // Handle completion
      console.log('AI completed:', message.content);
    }
  });

  return (
    <View style={styles.container}>
      <Text style={styles.title}>Solana AI Assistant</Text>
      
      {!connected && (
        <View style={styles.walletPrompt}>
          <Text>Connect your wallet to start chatting with the AI</Text>
          <ConnectWalletButton />
        </View>
      )}
      
      <FlatList
        data={messages}
        keyExtractor={(item) => item.id}
        renderItem={({ item }) => (
          <MessageBubble 
            message={item}
            isUser={item.role === 'user'}
          />
        )}
        style={styles.messagesList}
      />
      
      {currentOperation && (
        <View style={styles.operationStatus}>
          <ActivityIndicator size="small" />
          <Text style={styles.operationText}>{currentOperation}</Text>
        </View>
      )}
      
      <View style={styles.inputContainer}>
        <TextInput
          style={styles.textInput}
          value={input}
          onChangeText={handleInputChange}
          placeholder="Ask me about Solana operations..."
          multiline
          editable={!isLoading && connected}
        />
        <TouchableOpacity 
          style={[styles.sendButton, (!connected || isLoading) && styles.disabled]}
          onPress={handleSubmit}
          disabled={!connected || isLoading}
        >
          <Text style={styles.sendButtonText}>
            {isLoading ? "..." : "Send"}
          </Text>
        </TouchableOpacity>
      </View>
    </View>
  );
}

Chat Persistence & Management

The module provides comprehensive chat management capabilities:

chat_persistence

object

import { 
  saveChat, 
  fetchChat, 
  fetchChats, 
  deleteChat 
} from '@/modules/solana-agent-kit';

// Save chat to backend
const chatData = {
  id: generateUUID(),
  title: 'Portfolio Analysis Session',
  messages: messages,
  createdAt: new Date().toISOString(),
  updatedAt: new Date().toISOString()
};
await saveChat(chatData);

// Fetch specific chat
const chat = await fetchChat(chatId);

// Fetch all user chats
const userChats = await fetchChats();

// Delete chat
await deleteChat(chatId);

message_management

object

import { 
  saveMessages, 
  fetchMessages, 
  standardizeMessageFormat 
} from '@/modules/solana-agent-kit';

// Save messages with standardized format
const formattedMessages = messages.map(standardizeMessageFormat);
await saveMessages(chatId, formattedMessages);

// Fetch messages for a chat
const chatMessages = await fetchMessages(chatId);

// Convert for UI display
const uiMessages = convertToUIMessages(chatMessages);

user_management

object

import { createOrUpdateUser } from '@/modules/solana-agent-kit';

// Create or update user profile
await createOrUpdateUser({
  walletAddress: wallet.address,
  preferences: {
    aiModel: 'gpt-4o',
    language: 'en',
    notifications: true
  },
  lastActive: new Date().toISOString()
});

Advanced AI Features

Custom Tool Integration

function useCustomSolanaTools() {
  const { wallet } = useWallet();
  
  const customTools = useMemo(() => {
    if (!wallet) return [];
    
    return createVercelAITools({
      // Custom portfolio analysis tool
      analyzePortfolio: {
        description: 'Analyze user portfolio performance and provide insights',
        parameters: z.object({
          timeframe: z.enum(['1d', '7d', '30d', '90d', '1y']),
          includeNFTs: z.boolean().default(false)
        }),
        execute: async ({ timeframe, includeNFTs }) => {
          const analysis = await performPortfolioAnalysis(wallet.address, timeframe, includeNFTs);
          return {
            totalValue: analysis.totalValue,
            gainLoss: analysis.gainLoss,
            topPerformers: analysis.topPerformers,
            recommendations: analysis.recommendations
          };
        }
      },
      
      // Custom yield farming finder
      findYieldOpportunities: {
        description: 'Find the best yield farming opportunities for user tokens',
        parameters: z.object({
          minAPY: z.number().default(5),
          riskLevel: z.enum(['low', 'medium', 'high']).default('medium')
        }),
        execute: async ({ minAPY, riskLevel }) => {
          const opportunities = await findYieldFarmingOpps(wallet.address, minAPY, riskLevel);
          return opportunities;
        }
      }
    });
  }, [wallet]);
  
  return customTools;
}

Intelligent Transaction Batching

function useIntelligentBatching() {
  const { wallet, sendTransaction } = useWallet();
  
  const batchTransactions = async (operations) => {
    try {
      // AI analyzes operations and optimizes batching
      const optimizedBatches = await optimizeTransactionBatching(operations);
      
      const results = [];
      for (const batch of optimizedBatches) {
        const batchResult = await executeTransactionBatch(batch);
        results.push(batchResult);
      }
      
      return {
        success: true,
        results,
        totalGasSaved: calculateGasSavings(operations, optimizedBatches)
      };
    } catch (error) {
      console.error('Batch execution failed:', error);
      throw error;
    }
  };
  
  return { batchTransactions };
}

Smart Risk Assessment

function useRiskAssessment() {
  const assessTransactionRisk = async (transaction) => {
    try {
      const analysis = await analyzeTransactionRisk(transaction);
      
      return {
        riskLevel: analysis.riskLevel, // 'low', 'medium', 'high'
        factors: analysis.riskFactors,
        recommendations: analysis.recommendations,
        shouldProceed: analysis.riskLevel !== 'high'
      };
    } catch (error) {
      console.error('Risk assessment failed:', error);
      return { riskLevel: 'unknown', shouldProceed: false };
    }
  };
  
  return { assessTransactionRisk };
}

Integration with Other Modules

Embedded Wallets

Essential for transaction signing and wallet state management

Data Module

Real-time portfolio data and market information for AI analysis

All Trading Modules

AI can execute operations across Pump.fun, Raydium, Meteora, and more

Thread Module

AI can create posts about transactions and share insights

Unified AI Trading Assistant

import { useWallet } from '@/modules/wallet-providers';
import { useFetchTokens } from '@/modules/data-module';
import { useSwap } from '@/modules/swap';
import { usePumpFun } from '@/modules/pump-fun';
import { useChat } from '@/modules/solana-agent-kit';

function UnifiedTradingAssistant() {
  const { wallet } = useWallet();
  const { tokens, refetch } = useFetchTokens(wallet?.address);
  const { executeSwap } = useSwap();
  const { launchToken } = usePumpFun();

  const { messages, handleSubmit, input, handleInputChange } = useChat({
    id: 'trading-assistant',
    onToolCall: async (toolCall) => {
      switch (toolCall.toolName) {
        case 'executeSwap':
          const swapResult = await executeSwap(toolCall.args);
          await refetch(); // Refresh portfolio
          return swapResult;
          
        case 'launchToken':
          const launchResult = await launchToken(toolCall.args);
          await refetch(); // Refresh portfolio
          return launchResult;
          
        case 'getPortfolio':
          return tokens;
          
        default:
          throw new Error(`Unknown tool: ${toolCall.toolName}`);
      }
    }
  });

  return (
    <View style={styles.tradingAssistant}>
      <Text style={styles.title}>AI Trading Assistant</Text>
      <Text style={styles.subtitle}>
        Ask me to analyze your portfolio, execute trades, or launch tokens
      </Text>
      
      <ChatInterface 
        messages={messages}
        input={input}
        onInputChange={handleInputChange}
        onSubmit={handleSubmit}
        placeholder="Ask: 'What should I do with my portfolio?' or 'Swap 1 SOL for USDC'"
      />
    </View>
  );
}

Security & Best Practices

Transaction Approval: Always implement clear user confirmation for AI-initiated transactions. Never auto-execute without explicit user consent.

Context Management: The AI maintains conversation context, but be mindful of sensitive information in chat history.

function SecureAITransactions() {
  const { wallet } = useWallet();
  
  const confirmTransaction = async (transaction) => {
    return new Promise((resolve) => {
      Alert.alert(
        'Confirm Transaction',
        `The AI wants to execute a transaction:\n\n${formatTransactionDetails(transaction)}\n\nDo you approve?`,
        [
          { text: 'Cancel', onPress: () => resolve(false), style: 'cancel' },
          { text: 'Approve', onPress: () => resolve(true) }
        ]
      );
    });
  };
  
  const secureExecute = async (operation) => {
    try {
      // Risk assessment
      const riskAssessment = await assessOperationRisk(operation);
      
      if (riskAssessment.riskLevel === 'high') {
        Alert.alert('High Risk Operation', riskAssessment.warning);
        return;
      }
      
      // User confirmation
      const confirmed = await confirmTransaction(operation.transaction);
      if (!confirmed) {
        throw new Error('User cancelled operation');
      }
      
      // Execute with monitoring
      return await executeWithMonitoring(operation);
    } catch (error) {
      console.error('Secure execution failed:', error);
      throw error;
    }
  };
  
  return { secureExecute };
}

Error Handling & Troubleshooting

common_issues

object

Performance Considerations

Streaming Optimization: The AI uses streaming responses for better user experience. Ensure your UI handles partial updates smoothly.

Context Limits: Long conversations may hit AI context limits. Implement conversation summarization for extended chats.

API Reference

For detailed API documentation, see:

The Solana Agent Kit module transforms complex blockchain interactions into natural conversations, making Solana accessible to users of all technical levels while providing powerful automation capabilities for advanced users.

Introduction

Getting Started

Using LLMs

Modules

Server

Core Functionalities

AI Chat Interface

Solana Integration

Plugin Ecosystem

Smart Automation

Installation & Setup

Module Architecture

Core Hook: useChat

AI Model Configuration

Solana Agent Capabilities

Quick Start Examples

Chat Persistence & Management

Advanced AI Features

Custom Tool Integration

Intelligent Transaction Batching

Smart Risk Assessment

Integration with Other Modules

Embedded Wallets

Data Module

All Trading Modules

Thread Module

Unified AI Trading Assistant

Security & Best Practices

Error Handling & Troubleshooting

Performance Considerations

API Reference

Introduction

Getting Started

Using LLMs

Modules

Server

​Core Functionalities

AI Chat Interface

Solana Integration

Plugin Ecosystem

Smart Automation

​Installation & Setup

​Module Architecture

​Core Hook: useChat

​AI Model Configuration

​Solana Agent Capabilities

​Quick Start Examples

​Chat Persistence & Management

​Advanced AI Features

​Custom Tool Integration

​Intelligent Transaction Batching

​Smart Risk Assessment

​Integration with Other Modules

Embedded Wallets

Data Module

All Trading Modules

Thread Module

​Unified AI Trading Assistant

​Security & Best Practices

​Error Handling & Troubleshooting

​Performance Considerations

​API Reference

Core Functionalities

Installation & Setup

Module Architecture

Core Hook: useChat

AI Model Configuration

Solana Agent Capabilities

Quick Start Examples

Chat Persistence & Management

Advanced AI Features

Custom Tool Integration

Intelligent Transaction Batching

Smart Risk Assessment

Integration with Other Modules

Unified AI Trading Assistant

Security & Best Practices

Error Handling & Troubleshooting

Performance Considerations

API Reference