#AI编程
9513 related articles
Antigravity + Jules: How Autonomous AI Orchestration Enables Hands-Free Development Workflows
The open-source project antigravity-jules-orchestration combines Google Antigravity and Jules API via MCP integration to build an autonomous, hands-free AI development workflow. A deep dive into its architecture and value.
jules-app: A Self-Hosted Workbench for Google's Jules AI Agent
jules-app is an open-source, self-hosted workbench for Google's Jules AI coding agent, built with TypeScript. Explore its value, architecture, and what it reveals about the AI agent ecosystem.
Claude Code v2.1.270 Released: Fixes Regression Where Read-Only Git Commands Unexpectedly Prompted for Permission
Anthropic releases Claude Code v2.1.270, fixing a regression from v2.1.269 where read-only Git commands unexpectedly prompted for permission in long-running sessions.
GitHub Project Overview: prateek0221/jules — Critically Lacking Information
The GitHub repo prateek0221/jules has only 1 star, no code, and no README. We review available info and provide background on Google's Jules AI coding agent.
google-jules-mcp: An MCP Server Bridging Google Jules Coding Agent
google-jules-mcp is an open-source MCP server that wraps Google Jules AI coding agent's API into a standard Model Context Protocol interface for plug-and-play use.
autonomous-agents: A Collection of Autonomous AI Agents for Codebase Optimization
autonomous-agents is an open-source collection of autonomous AI agents for systematic codebase optimization, inspired by Google Jules, using specialized agents for performance, security, UX, and testing.
google-jules-mcp: A MCP Bridge Connecting Claude to Google Jules
google-jules-mcp is a dedicated MCP server that bridges AI clients like Claude and Cursor with Google Jules Agentic API, supporting code generation, plan approval, and session management.
jules-skill: Control Google's AI Coding Agent Jules from the Command Line
jules-skill is an open-source CLI tool that manages Google's AI coding agent Jules via REST API, supporting Claude Code, Gemini CLI, and terminal workflows.
WPF Recreation of Google Jules: A Desktop Exploration of AI Agents
The open-source project danyooooo/jules recreates Google's AI Agent Jules using WPF and C#, exploring how to bring AI Agent capabilities to the Windows desktop. This article breaks down the tech choices, value, and limitations.
Running Claude Code Locally: A Fully Offline AI Coding Solution for Apple Silicon
claude-code-local runs Claude Code 100% offline on Apple Silicon via MLX, supporting Qwen 3.5 122B at 65 tok/s — built for NDA, legal, and medical privacy workflows.
Kimi 2.6 Code: A Terminal-Native Coding Agent Built on the Moonshot Model
kimi-2-6-code is a terminal-native coding agent built in TypeScript on Moonshot's Kimi K2.6 model. Learn about its design, model choice, and maturity level.
bert4torch: A Deep Dive into an Elegant PyTorch Transformers Implementation
bert4torch is an elegant PyTorch Transformers library with 1,334 GitHub Stars. This article analyzes its technical positioning, use cases, and how it differs from Hugging Face.
Building a Self-Hosted Research AI Agent: A Tool Selection Guide Without Paid APIs
How to build a self-hosted research AI Agent without paid APIs or cloud services? This guide covers SearXNG, Crawl4AI, Firecrawl, anti-bot strategies, and core Agent skill design.
AI Can Write Code — So Why Do Software Projects Keep Failing?
Paul Ford argues in the NYT that AI writes decent code but makes it easier to do bad work. Why engineering judgment and collaboration matter more than ever.
QApilot MCP: Test Android Apps in Natural Language Inside Your AI Coding Assistant
QApilot MCP lets developers test Android apps in natural language inside Claude, Cursor, and Codex — no Appium code needed. Auto-generates reusable Gherkin test cases.
ajisai: A Preset Management Tool for Unifying Rules and Prompts Across AI Coding Assistants
ajisai is a Go-based preset manager for AI coding assistants like Cursor and Claude Code. Package rules and prompts once, deploy everywhere. Early-stage but promising.
Cortex: Convert API Specs into Docs, SDKs, and MCP Servers in One Click
Cortex is an open-source tool that converts OpenAPI, GraphQL, gRPC and more into interactive docs, typed SDKs in 11 languages, and MCP servers for AI agents.
Deep Dog 2: The Open-Source Research Agent Ranked #5 on DeepResearch Bench
Deep Dog 2 is an open-source deep research agent ranked #5 on DeepResearch Bench, featuring one-line install, DeepSeek V4 Flash + Exa, and specialized subagents for PubMed, Reddit, SEC Edgar, and Arxiv. MIT licensed and free.
Code with Claude Event: AI-Powered Creative Experiments on Tiny Computers
At Code with Claude, Anthropic gave participants tiny computers to build creative projects with AI assistance, showcasing how AI coding tools empower the maker community.
Claude Code Adds Agent View: A Research Preview for Unified Session Management
Claude Code's new Agent View feature (research preview) consolidates all coding sessions into a unified list, advancing AI tools toward multi-agent orchestration.