2026GitHubAI Agent

2026AI AgentRAGGitHub Trending8

1. RuView — WiFi

# RuViewimport numpy as npfrom scipy import signalclass RuViewProcessor: def __init__(self, num_antennas=3): self.num_antennas = num_antennas self.csi_buffer = [] def extract_csi_features(self, csi_data): """WiFi""" # 1. calibrated = self.calibrate_phase(csi_data) # 2. doppler = self.compute_doppler_shift(calibrated) breathing = self.extract_breathing_pattern(calibrated) # 3. pose = self.classify_pose(doppler, breathing) return { 'pose': pose, 'heart_rate': self.estimate_heart_rate(breathing), 'presence': self.detect_presence(doppler) } def calibrate_phase(self, raw_csi): """""" return signal.detrend(raw_csi, axis=0)

****

-

**** Rust + Python

2. RAG-Anything — RAG

**** All-in-OneRAG

# RAG-Anythingfrom rag_anything import RAGPipeline, DocumentLoader, VectorStore# 1. loader = DocumentLoader( supported_formats=['pdf', 'docx', 'md', 'html', 'txt'], chunk_size=512, chunk_overlap=50)# 2. from rag_anything.chunking import SemanticChunkerchunker = SemanticChunker( embedding_model='bge-large-zh-v1.5', threshold=0.7 # )# 3. vector_store = VectorStore( backend='milvus', # Milvus/Qdrant/Chroma dimension=1024, metric='cosine')# 4. Pipelinepipeline = RAGPipeline( loader=loader, chunker=chunker, embedder='bge-large-zh-v1.5', vector_store=vector_store, llm='qwen-max' # LLM)# 5. result = pipeline.query( "Explain the attention mechanism in transformers", top_k=5, rerank=True # Reranker)

****

• Reranker
• MCPClaude

3. claude-context —

MCP Claude

// MCP Serverimport { Server } from '@anthropic-ai/mcp-server';const server = new Server({ name: 'claude-context', version: '1.0.0'});// 1: server.tool('search_codebase', { query: { type: 'string', description: '' }, path: { type: 'string', description: '' }}, async ({ query, path }) => { // + const results = await vectorSearch(query, path); return { content: results };});// 2: server.tool('analyze_structure', { filePath: { type: 'string' }}, async ({ filePath }) => { const ast = parseCodeToAST(filePath); return { functions: ast.functions, classes: ast.classes, imports: ast.imports };});server.start();

• Claude
• token

4. FinceptTerminal — AI

# FinceptTerminalclass FinceptTerminal: def __init__(self): self.data_provider = MultiSourceDataProvider([ AlphaVantageProvider(), YahooFinanceProvider(), BinanceProvider() ]) self.analysis_engine = AIAnalysisEngine( model='gpt-4-turbo', tools=[TechnicalAnalysis(), FundamentalAnalysis()] ) self.visualizer = ChartVisualizer() def analyze_stock(self, symbol: str): # 1. market_data = self.data_provider.get_market_data(symbol) news = self.data_provider.get_news(symbol) financials = self.data_provider.get_financials(symbol) # 2. AI analysis = self.analysis_engine.analyze({ 'price_history': market_data, 'news_sentiment': news, 'financial_health': financials }) # 3. charts = self.visualizer.create_dashboard([ ('price', market_data), ('indicators', analysis.technical), ('sentiment', analysis.sentiment) ]) return { 'analysis': analysis, 'charts': charts, 'recommendation': analysis.recommendation }

****

- AI

• Python

5. ai-agents-for-beginners —

12

# Agentfrom semantic_kernel import Kernelfrom semantic_kernel.connectors.ai.open_ai import AzureChatCompletionfrom semantic_kernel.planning import SequentialPlanner# 1. Kernelkernel = Kernel()kernel.add_service(AzureChatCompletion())# 2. Pluginclass WeatherPlugin: @kernel.function() def get_weather(self, city: str) -> str: # API return f"{city}25°C"# 3. kernel.add_plugin(WeatherPlugin(), plugin_name="weather")# 4. planner = SequentialPlanner(kernel)# 5. result = await planner.invoke("")

1AI Agent

2024: Agent2025: Agent2026: Agent- MCPModel Context Protocol- Agent- 

2

3RAG

RAG: : Unstructured + LangChain : BGE/Jina/M3E : Milvus/Qdrant/Chroma : BM25 + + Rerank : Claude/GPT/Qwen

4AI

• ****FinceptTerminalQuantitative Agent
• ****Agent

• ---

• ****Agent

5

Thunderbolt "AI You Control"

// Thunderboltinterface ThunderboltConfig { modelRegistry: ModelRegistry; // dataGovernance: DataPolicy; // localFirst: boolean; // noVendorLock: boolean; // }class Thunderbolt { async chat(message: string) { // const model = this.config.modelRegistry.selectByTask(message); // if (this.config.localFirst) { return this.localModel.inference(message); } return model.inference(message); }}

1. ** **Star
2. ** **AI Agent
3. ** **RAG-AnythingTrendRadar

4. ---

2026AI""""RAG-AnythingRAGRuViewclaude-context

AI****