OpenAI 发布了一项新功能 “深度研究”(Deep Research),通过长逻辑链让 ChatGPT 处理复杂任务。Jina AI 认为,这项功能的核心原理本质上是通过 while 循环实现的 “搜索 + 访问 + 反思” 的流程。为了验证这一观点,他们结合 Jina AI(一款集成 LLM 的搜索工具)与 Gemini Flash,成功复现了与 OpenAI Deep Research 类似的功能。

流程图

如何通过 Gemini 实现 OpenAI DeepResearch 功能

点击放大

安装

export GEMINI_API_KEY=...  # for gemini api, ask han
export JINA_API_KEY=jina_...  # free jina api key, get from https://jina.ai/reader
export BRAVE_API_KEY=...  # (optional, when not given it uses duckduckgo) brave search provide free key, ask han

git clone https://github.com/jina-ai/node-DeepResearch.git
cd node-DeepResearch
npm install 

示例

# example: no tool calling 
npm run dev "1+1="
npm run dev "what is the capital of France?"

# example: 2-step
npm run dev "what is the latest news from Jina AI?"

# example: 3-step
npm run dev "what is the twitter account of jina ai's founder"

# example: 13-step, ambiguious question (no def of "big")
npm run dev "who is bigger? cohere, jina ai, voyage?"

# example: open question, research-like, long chain of thoughts
npm run dev "who will be president of US in 2028?"
npm run dev "what should be jina ai strategy for 2025?"

获取 Jina AI 最新博客文章

整个流程被分为三步:访问 Jina AI 网站的博客页面、阅读内容、找到最新博客文章。

对 Cohere、Jina AI、Voyage AI 进行对比分析

通过搜索、访问、反思、回答,四步的循环操作(一共进行了13步操作)对 Cohere、Jina AI、Voyage AI 三款 AI 工具进行对比分析。

项目地址:https://github.com/jina-ai/node-DeepResearch