用Stagehand与DeepAgents构建浏览器Agent实战指南
last night i spent some time building a browser agent with stagehand from @brows…
做Agent的同学必看，这是一份可复现的浏览器Agent实战记录，从一次性构建到基于trace迭代优化的完整流程，赶紧照着指南搭一个试试。
last night i spent some time building a browser agent with stagehand from @browserbase and deepagents from @LangChain
昨晚我花了一些时间，用@browserbase的stagehand和@LangChain的deepagents构建了一个浏览器代理。
i wanted to build something that required pretty complex web navigation to prove the value of agentic browser use!
我想构建一个需要相当复杂的网页导航的东西，以证明代理式浏览器使用的价值！
my agent plays a game called "map tap" where you are assigned a city around the globe, and have to drag the map, zoom in, and click as close to the given city as possible. you get points based on how close you are.
我的代理玩一个叫“地图点击”的游戏，游戏会给你分配一个全球各地的城市，你必须拖动地图、放大，并尽可能点击接近给定城市的位置。根据你点击的接近程度得分。
1. one shotted the agent w/ browserbase + langchain docs, it scored ~300/1000 points 2. asked my coding agent to review the trace and make it faster (smaller model) and more accurate (higher score) 3. came back 10 minutes later to see a perfect score screenshot in my trace
1. 用browserbase + langchain文档一次性完成代理，得分约300/1000分 2. 让我的编码代理审查轨迹并使其更快（更小模型）和更准确（更高分数） 3. 10分钟后回来，在我的轨迹中看到完美得分的截图。
here's a guide on how to build your own browser agent! https://docs.langchain.com/oss/python/integrations/tools/stagehand
这里有一份关于如何构建你自己的浏览器代理的指南！ https://docs.langchain.com/oss/python/integrations/tools/stagehand
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力