Solana Agent Kit provides integration with Elfa AI for retrieving social media intelligence, market data, and trend analysis from platforms like Twitter. This documentation covers the available functions and tools for interacting with Elfa AI API.
Overview
Elfa AI integration enables access to:
- Smart mentions detection
- Top mentions by ticker symbol
- Keyword-based mentions search
- Trending tokens analysis
- Twitter account statistics
- API key status monitoring
Configuration
Before using Elfa AI integration, you need to configure your API key in the Solana Agent Kit:
import { SolanaAgentKit } from "solana-agent-kit";
const agent = new SolanaAgentKit(
privateKey,
rpcUrl,
{
ELFA_AI_API_KEY: "your_elfa_ai_api_key"
}
);
Core Functions
The integration provides several core functions for interacting with the Elfa AI API:
API Health Check
const pingResult = await agent.methods.pingElfaAiApi();
// Returns: { message: "pong" }
API Key Status
const keyStatus = await agent.methods.getElfaAiApiKeyStatus();
// Returns detailed information about API key status including usage limits
Smart Mentions
Retrieve tweets from smart accounts with high engagement:
const smartMentions = await agent.methods.getSmartMentions(
100, // limit
0 // offset
);
Example response:
{
"success": true,
"data": [
{
"id": "611245869",
"type": "post",
"content": "In my opinion, it's a great time to add $ETH.",
"originalUrl": "/EricTrump/status/1886541132903133230",
"likeCount": 48036,
"quoteCount": 3103,
"replyCount": 8900,
"repostCount": 7981,
"viewCount": 5660995,
"mentionedAt": "2025-02-03T22:23:50.000Z",
"account": {
"id": 83583,
"username": "EricTrump",
"followerCount": 5500559,
"isVerified": true
}
}
],
"metadata": {
"total": 6,
"limit": 100,
"offset": 0
}
}
Top Mentions by Ticker
Retrieve top tweets for a specific ticker symbol:
const topMentions = await agent.methods.getTopMentionsByTicker(
"SOL", // ticker
"1h", // timeWindow
1, // page
10, // pageSize
false // includeAccountDetails
);
Example response:
{
"success": true,
"data": {
"data": [
{
"id": 612200471,
"twitter_id": "1886663937518645714",
"content": "Same story for $SOL - looks like a failed breakdown.",
"mentioned_at": "2025-02-04T06:31:49+00:00",
"type": "post",
"metrics": {
"like_count": 45,
"reply_count": 6,
"repost_count": 7,
"view_count": 1744
}
}
],
"total": 12,
"page": 1,
"pageSize": 10
}
}
Search Mentions by Keywords
Search for tweets containing specific keywords within a date range:
const searchResults = await agent.methods.searchMentionsByKeywords(
"ai, agent", // comma-separated keywords
1622505600, // from timestamp
1625097600, // to timestamp
20 // limit
);
Example response:
{
"success": true,
"data": [
{
"id": 612258820,
"twitter_id": "1886671035048845535",
"content": "The Move AI Hackathon 🛠️\n\nUnlock limitless possibilities...",
"mentioned_at": "2025-02-04T07:00:02+00:00",
"type": "quote",
"metrics": {
"like_count": 1,
"reply_count": 0,
"repost_count": 0,
"view_count": 0
}
}
],
"metadata": {
"total": 1875,
"cursor": "FGluY2x1ZGVfY29udGV4d..."
}
}
Trending Tokens
Get tokens trending in social media discussions:
const trendingTokens = await agent.methods.getTrendingTokens();
Example response:
{
"success": true,
"data": {
"total": 5,
"page": 1,
"pageSize": 5,
"data": [
{
"token": "eth",
"current_count": 916,
"previous_count": 377,
"change_percent": 142.97
},
{
"token": "btc",
"current_count": 580,
"previous_count": 458,
"change_percent": 26.64
}
]
}
}
Get engagement and influence metrics for a Twitter account:
const accountStats = await agent.methods.getSmartTwitterAccountStats("elonmusk");
Example response:
{
"success": true,
"data": {
"smartFollowingCount": 5913,
"averageEngagement": 30714784.98833819,
"followerEngagementRatio": 0.1423006675490259
}
}
For AI agent integration with LangChain, the following tools are available:
import { ElfaPingTool } from "solana-agent-kit";
const pingTool = new ElfaPingTool(agent);
import { ElfaApiKeyStatusTool } from "solana-agent-kit";
const apiKeyStatusTool = new ElfaApiKeyStatusTool(agent);
import { ElfaGetMentionsTool } from "solana-agent-kit";
const smartMentionsTool = new ElfaGetMentionsTool(agent);
// Tool input (JSON string):
const input = JSON.stringify({
limit: 100,
offset: 0
});
import { ElfaGetTopMentionsTool } from "solana-agent-kit";
const topMentionsTool = new ElfaGetTopMentionsTool(agent);
// Tool input (JSON string):
const input = JSON.stringify({
ticker: "SOL",
timeWindow: "1h",
page: 1,
pageSize: 10,
includeAccountDetails: false
});
import { ElfaSearchMentionsTool } from "solana-agent-kit";
const searchMentionsTool = new ElfaSearchMentionsTool(agent);
// Tool input (JSON string):
const input = JSON.stringify({
keywords: "ai, agent",
from: 1622505600,
to: 1625097600,
limit: 20
});
import { ElfaTrendingTokensTool } from "solana-agent-kit";
const trendingTokensTool = new ElfaTrendingTokensTool(agent);
import { ElfaAccountSmartStatsTool } from "solana-agent-kit";
const accountStatsTool = new ElfaAccountSmartStatsTool(agent);
// Tool input: username as string
const input = "elonmusk";
Action Definitions
In addition to functions and tools, Solana Agent Kit provides actions for interacting with Elfa AI:
Ping Action
const pingResult = await agent.methods.executeAction("ELFA_PING_ACTION", {});
API Key Status Action
const keyStatus = await agent.methods.executeAction("ELFA_API_KEY_STATUS_ACTION", {});
Smart Mentions Action
const smartMentions = await agent.methods.executeAction("ELFA_GET_SMART_MENTIONS_ACTION", {
limit: 100,
offset: 0
});
Top Mentions Action
const topMentions = await agent.methods.executeAction("ELFA_GET_TOP_MENTIONS_BY_TICKER_ACTION", {
ticker: "SOL",
timeWindow: "1h",
page: 1,
pageSize: 10,
includeAccountDetails: false
});
Search Mentions Action
const searchResults = await agent.methods.executeAction("ELFA_SEARCH_MENTIONS_BY_KEYWORDS_ACTION", {
keywords: "ai, agent",
from: 1622505600,
to: 1625097600,
limit: 20
});
Trending Tokens Action
const trendingTokens = await agent.methods.executeAction("ELFA_TRENDING_TOKENS_ACTION", {});
Account Smart Stats Action
const accountStats = await agent.methods.executeAction("ELFA_SMART_TWITTER_ACCOUNT_STATS_ACTION", {
username: "elonmusk"
});
Parameter Details
Time Windows
The following time windows are supported for various endpoints:
1h - Last hour24h - Last 24 hours7d - Last 7 days14d - Last 14 days30d - Last 30 days
Most endpoints support pagination with the following parameters:
limit or pageSize - Number of results per pageoffset or page - Page number or offset for resultscursor - For cursor-based pagination (used in search endpoints)
Error Handling
All functions and tools include error handling. Here’s an example of handling errors:
try {
const result = await agent.methods.getTopMentionsByTicker("SOL");
if (!result.success) {
console.error("API error:", result.error);
// Handle error
} else {
console.log("Top mentions:", result.data);
// Process data
}
} catch (error) {
console.error("Exception:", error.message);
// Handle exception
}
Utility Functions
The integration includes several utility functions:
Create Axios Instance
function createAxiosInstance(apiKey: string): AxiosInstance {
return axios.create({
baseURL: "https://api.elfa.ai",
headers: {
"x-elfa-api-key": apiKey,
"Content-Type": "application/json",
},
});
}
Integration with AI Agents
For AI agents built with LangChain, you can register these tools:
import {
ElfaPingTool,
ElfaGetTopMentionsTool,
ElfaTrendingTokensTool
} from "solana-agent-kit";
import { OpenAI } from "langchain/llms/openai";
import { initializeAgentExecutorWithOptions } from "langchain/agents";
// Create LangChain tools
const tools = [
new ElfaPingTool(agent),
new ElfaGetTopMentionsTool(agent),
new ElfaTrendingTokensTool(agent)
];
// Create LLM
const model = new OpenAI({ temperature: 0 });
// Create agent
const executor = await initializeAgentExecutorWithOptions(
tools,
model,
{
agentType: "structured-chat-zero-shot-react-description",
}
);
// Execute agent
const result = await executor.call({
input: "What are the trending tokens today?"
});
Best Practices
- Rate Limiting: Monitor your API usage through the key status endpoint
- Data Freshness: Social media data can become stale quickly, consider time windows carefully
- Error Handling: Implement comprehensive error handling for all API calls
- Pagination: For large datasets, implement pagination to avoid timeouts
- Date Ranges: For search queries, use reasonable date ranges to optimize performance
Resources
