 

外包：3DGS交互Demo开发需求

2025-12-17 分类：前沿哨所阅读(2) 评论(0) 赞(0)

智谱 GLM，支持多语言、多任务推理。从写作到代码生成，从搜索到知识问答，AI 生产力的中国解法。

本文介绍了一个前沿技术外包需求，旨在开发手机端适配的Web 3D页面，核心功能是加载并展示3D Gaussian Splatting模型。技术栈要求Three.js和@lumaai/luma-web，支持加载.splat或.ply格式模型（约30-50MB），实现手指旋转、缩放、惯性滑动交互，并通过Raycasting实现点击模型特定部位弹出HTML信息层。需要针对移动端进行性能优化，确保不闪退、不卡顿。交付物包括完整前端源码、换模型文档、简易部署文档，以及模型压缩指导。预算8000元，工期10天。此项目展示了前沿3D渲染技术在Web端的应用，对AI和GPU领域有重要价值。

原文链接：V2EX 分享发现

赞(0)

未经允许不得转载：Toy Tech Blog » 外包：3DGS交互Demo开发需求

分享到

评论抢沙发

快讯

AI Agent Development in Practice: Building an Interactive Problem-Solving Visualization Platform

This article introduces an open-source project called ThinkingMap, which visualizes the AI problem-solving process based on multi-agent collaboration technology. The project utilizes backend technologies such as Go, Eino, Redis, and PostgreSQL, along with frontend technologies including Next.js, ReactFlow, Shadcn/ui, and Zustand to implement complete Agent orchestration, streaming output processing, and context logic management. The author provides a detailed record of the entire development process, including architectural design, Agent development approach, and practical implementation experience, offering valuable insights for AI developers. The project is now open-source with GitHub links and a demo address available, supporting real-time user intervention and collaboration, making it ideal for learning practical AI Agent applications.

Original Link:V2EX Share & Discover

52分钟前
Open Source Chrome Extension Gemini Nexus: One-Click Access to Gemini 3 Flash in Sidebar

This article introduces Gemini Nexus, an open-source Chrome extension that allows users to directly access Google Gemini's official Gemini 3 Flash model from their browser sidebar. Users simply need to log in to the official Gemini website to use this extension and access the fast AI model, enabling convenient text generation and interactive features. The extension is lightweight at just 53.0 KB, easy to use, and requires no additional configuration. Developed by the open-source community, this tool demonstrates the growing integration of AI technology in everyday applications, providing tech enthusiasts with an efficient AI-powered assistant solution. Meanwhile, it has sparked discussions in the Linux.do community, showcasing the vitality of the open-source ecosystem in AI tool innovation.

Original Link:Linux.do

52分钟前
Gemini CLI API Configuration: .env File Not Working, Environment Variables Work

On the Linux.do forum, a user reported a Gemini CLI API configuration issue: API requests work normally when configured with environment variables, but fail when using a .env file. This topic attracted 11 posts from 5 participants, delving into configuration differences and potential solutions. For developers relying on Gemini CLI, this issue provides practical technical insights, revealing common pitfalls in configuration tools and emphasizing the critical role of environment variables. The article not only resolves specific configuration challenges but also helps optimize API settings and improve development efficiency, serving as a valuable reference for AI tool users.

Original Link:Linux.do

52分钟前
Gemini Network Switch Causes Performance Decline as Users Seek Optimization Solutions

A user on the Linux.do forum reported performance issues while using the Gemini AI model: after switching network nodes from a clean US residential connection to a clean Japanese residential connection, the model's reasoning capabilities significantly weakened, with outputs becoming chaotic or terminating midway. The user described this as a 'degradation in intelligence' phenomenon and asked if there was a way to resolve it. This reflects stability issues with AI services under different network conditions, potentially involving network latency, data transmission efficiency, or server load optimization. For users relying on AI technology, this issue impacts work efficiency, highlighting the critical influence of network environment on AI responsiveness. The post has sparked discussions, indicating a common pain point that requires technical support and calling on AI service providers to strengthen cross-regional performance stability.

Original Link:Linux.do

53分钟前
AI Agent开发实战：构建交互式问题解决可视化平台

本文介绍了一个名为ThinkingMap的开源项目，它基于多智能体协作技术，将AI问题解决过程可视化。项目使用Go、Eino、Redis、PostgreSQL等后端技术和Next.js、ReactFlow、Shadcn/ui、Zustand等前端技术，实现了完整的Agent编排、流式输出处理和上下文逻辑管理。作者详细记录了开发全过程，包括架构设计、Agent开发思路和落地经验，为AI开发者提供宝贵参考。项目已开源，GitHub链接和体验地址可用，支持用户实时干预和协作，适合学习AI Agent的实际应用。

原文链接：V2EX 分享发现

53分钟前
开源Chrome插件Gemini Nexus：侧边栏一键调用Gemini 3 Flash

本文介绍了一个名为Gemini Nexus的开源Chrome插件，它允许用户在浏览器侧边栏直接调用Google Gemini官网的Gemini 3 Flash模型。用户只需登录Gemini官网，即可使用该插件访问快速模式AI模型，实现便捷的文本生成和交互功能。插件体积小巧，仅53.0 KB，操作简单，无需额外配置。该工具由开源社区开发，体现了AI技术在日常应用中的普及，为科技爱好者提供了高效的AI辅助解决方案。同时，它在Linux.do社区引发讨论，展示了开源生态在AI工具创新中的活力。

原文链接：Linux.do

53分钟前