ai-trend-agent

Autonomous Multi-Agent AI Trend Intelligence Platform

---

Overview

AI evolves rapidly. Hundreds of papers and repositories appear daily — but identifying true emerging trends requires:

• Cross-source signal aggregation
• Semantic similarity analysis
• Velocity-based ranking
• Multi-agent orchestration

ai-trend-agent is an open-source autonomous multi-agent system that monitors emerging AI research and GitHub repositories to detect meaningful trends in artificial intelligence.

It ingests raw signals, embeds and stores them in a vector database, clusters recent activity, detects emerging themes, and generates structured intelligence reports — powered by open-source tools and local LLMs.

---

Tech Stack (Fully Open Source)

• Python
• LangGraph / LangChain
• Ollama (local LLM + embeddings)
• Qdrant (vector database)
• FastAPI
• Docker (Docker desktop)

---

Data Sources

• RSS feeds: ArXiv (cs.AI), OpenAI News
• GitHub (AI/LLM/Agent repositories)

All data is normalized into a unified schema: { "id": "...", "source": "arxiv | github | openai", "timestamp": "ISO8601", "content": "...", "vector": [...], "metadata": {...} }

---

Run Locally

• docker compose up --build (start services: main app, qdrant)
• ollama run llama3.1:8b (ensure ollama is running: access /llm-test endpoint)
• ollama stop llama3.1:8b (free resources when finished)

Development Environment

• Windows 11 Home
• Visual Studio Code with Copilot
• LLM on host for easy access to GPU and everything else on docker
• Resources: 16G RAM, 12G GPU, 6-core CPU