Current prototype:https://etf-navigator-f8d67fe2.base44.app/
🎯 Overview
ETF Navigator is an AI-powered investment research platform that helps users discover, analyze, and monitor sustainable ETFs (e.g., Clean Energy, ESG-themed funds).
Unlike traditional financial dashboards, our system is built around an AI Agent that:
Fetches and synthesizes real-time market data Summarizes news and trends in plain language Highlights sustainability-aligned investment opportunities Supports both U.S. and Chinese markets This project directly responds to the track theme: “AI Agents Unlock Finance for Sustainability”.
🤖 AI Agent Design
The core of ETF Navigator is an autonomous AI Agent implemented via Base44’s LLM integration layer. It operates as follows:
Perception: Receives user queries or scheduled triggers (e.g., “Show today’s market update”).
Reasoning & Data Retrieval: Uses structured prompts to invoke a large language model (LLM) with internet access (add_context_from_internet: true). The LLM is instructed to:
Search for credible financial sources (e.g., AP News, official fund websites) Extract structured data (indices, ETF prices, sector performance) Identify sustainability-relevant categories (e.g., “Clean Energy”, “Low Carbon”) Action & Transparency: Returns validated JSON responses that power the UI. All recommendations include data provenance (e.g., source URLs) to ensure auditability.
🔧 Technical Implementation
Component Technology / Approach Frontend & Agent Logic Built on Base44 – a low-code platform for AI-native apps Data Sourcing Public financial data retrieved via LLM with internet access Sustainability Focus Dedicated sections for "Clean Energy ETFs", ESG analysis, and China's green finance Ethical Design - No user financial data collected
- All outputs labeled “for informational purposes only”
- Sources disclosed in footnotes 🔐 Note: The live demo is hosted on Base44 for rapid prototyping. All custom logic (page components, prompts, data models) has been manually exported and is included in this repository for transparency and review. 📁 Repository Structure
/pages – Core page components (Dashboard, Research, Watchlist, etc.) /components – Reusable UI modules (MarketOverview, PopularETFs, etc.) /entities – Data models (ETF, WatchlistItem) /integrations – Core LLM invocation logic (Core.js) proposal.pdf – Full Stage 1 proposal demo.mp4 – 3-minute demonstration video 📜 Submission Compliance
This repository is private and shared with NYUSH_DIC and team members as requested. © [DIC04-AI_Finance], 2025. For DIC 2025 competition use only.
README-ZH
ETF Navigator:面向可持续投资发现的 AI Agent
DIC 2025 金融赛道(赛道 04)– 第一赛段提交 团队:DIC04-AI_Finance 代码仓库:DIC04-AI_Finance 在线演示:暂定的演示地址:https://etf-navigator-f8d67fe2.base44.app/ 🎯 项目概述
ETF Navigator 是一个由 AI 驱动的投资研究平台,帮助用户发现、分析和监控可持续型 ETF(例如清洁能源、ESG 主题基金)。
与传统金融仪表盘不同,本系统围绕一个AI Agent构建,具备以下能力:
获取并整合实时市场数据 用通俗语言总结新闻与趋势 突出展示与可持续发展目标对齐的投资机会 同时支持美国与中国市场 本项目直接响应赛道主题: “AI Agent 解锁可持续发展的金融潜力”。
🤖 AI Agent 设计
ETF Navigator 的核心是一个自主运行的 AI Agent,通过 Base44 平台的 LLM 集成层实现,其工作流程如下:
感知(Perception) 接收用户查询或定时触发指令(例如:“显示今日市场简报”)。
推理与数据获取(Reasoning & Data Retrieval) 使用结构化提示(prompt)调用具备联网能力的大语言模型(LLM,启用 add_context_from_internet: true),指令 LLM:
从可信金融信源(如美联社 AP News、基金公司官网)搜索信息 提取结构化数据(指数、ETF 价格、板块表现) 识别与可持续性相关的类别(如“清洁能源”、“低碳”) 行动与透明度(Action & Transparency) 返回经校验的 JSON 数据驱动前端界面。 所有推荐均包含数据来源引用(如源链接),确保可审计性与透明度。
🔧 技术实现
组件 技术与方法 前端与 Agent 逻辑 基于 Base44 构建 —— 一个面向 AI 原生应用的低代码平台 数据获取 通过 LLM 联网获取公开金融数据 可持续聚焦 专设“清洁能源 ETF”、“ESG 分析”、“中国绿色金融”等模块 伦理设计 - 不收集用户任何财务或身份信息
- 所有输出标注“仅限信息参考,不构成投资建议”
- 在脚注中披露数据来源 🔐 说明:在线演示托管于 Base44,用于快速原型开发。所有自定义逻辑(页面组件、提示词、数据模型)均已手动导出,并完整包含在本代码仓库中,以确保透明度与评审可查性。 📁 代码仓库结构
/pages – 核心页面组件(Dashboard、Research、Watchlist 等) /components – 可复用 UI 模块(MarketOverview、PopularETFs 等) /entities – 数据模型(ETF、WatchlistItem) /integrations – 核心 LLM 调用逻辑(Core.js) proposal.pdf – 完整第一赛段提案 demo.mp4 – 3 分钟演示视频 📜 提交合规性声明
本仓库为私有仓库,并已按要求邀请 NYUSH_DIC 与其他参赛队员为协作者。 © DIC04-AI_Finance,2025。仅供 DIC 2025 竞赛使用。
Built With
- base44
- llm
Log in or sign up for Devpost to join the conversation.