Skip to main content
Back

Google Chrome launches DevTools MCP: AI agents debug live pages

Meta’s open-source code world model, Claude 4 in Copilot, Gemini voice, GPT-5 Codex in Cursor, Colab adds freeze runtime

September 25, 2025 4 min read

Hey

Welcome to AlphaSignal, the most read source of news by AI engineers and researchers.


Every day, we identify and summarize the top 1% of news, papers, models, and repos, so you’re always up to date.


Here’s today’s roundup:

Summary

Read time: 4 min 23 sec

Top News

▸ Google releases new Chrome MCP server, giving AI agents browser debugging

Warp

▸ Outperform Claude Code and Cursor with Warp’s prompt-to-code precision

Signals

▸ Meta debuts open-source world model simulating Python and Bash execution

▸ Microsoft adds Claude 4 models to Copilot for enterprise users

▸ Google introduces native audio for more natural Gemini voice conversations

▸ Cursor integrates GPT-5-Codex to its platform

▸ Google Colab now lets you freeze runtimes for reproducible results

Trending Reads

▸ Google reports 90% of developers now use AI daily

▸ How far can coding agents go improving themselves at runtime?

▸ Google Cloud details tools to build and scale AI agents

How To

▸ Build code from Figma designs with GPT-5-Codex in Cursor

Top News

Google Chrome introduces DevTools MCP, enabling AI agents to debug and analyze live web pages

alpha_signal_image_2

Google has launched the Chrome DevTools MCP server in public preview, letting AI coding agents directly interact with a live Chrome browser.

Instead of blindly generating code, agents can now see, test, and debug in real-time with full DevTools capabilities.


Key Features

  • AI agents can now see and test code directly in Chrome.

  • 26 built-in tools for debugging, performance tracing, and network analysis.

  • Agents can simulate real user actions like clicks and form fills.

  • Supports automated performance audits with trace data.

  • Runs securely in sandboxed Chrome environments.

How It Works

  • Connects your AI agent to Chrome via the open-source MCP.

  • Agents can issue commands like performance_start_trace to collect live performance data.

  • Results feed back into the agent, allowing it to analyze errors, verify fixes, and suggest optimizations.

  • Works across AI assistants like Claude, Cursor, Copilot, or Gemini CLI.

Access & Availability

  • Public preview available today on GitHub.

  • Install via npx chrome-devtools-mcp@latest.

  • Configurable in any MCP-compatible client.

Try now
alpha_signal_image_3

Warp Code Launches, Saving Developers 1–2 Hours Daily

Warp just launched Warp Code and early users report saving 1–2 hours per day. “Almost right” agent written code leaves developers debugging instead of shipping.

Warp provides prompt-to-code precision from one platform.

  • Early users generate 150M+ lines weekly with 97% diff code acceptance

  • Outperforms Claude Code and Cursor on coding benchmarks

  • Easy code review and edit without leaving Warp

  • Use code ALPHASIGNAL for one month of Warp Pro for $1

Download warp

partner with us

Signals

Meta releases Code World Model, a 32B open-source LLM for code execution research

3,721 Likes

Microsoft expands Copilot beyond OpenAI, letting enterprises switch to Claude reasoning models

3,771 Likes

Google enhances Live API to ignore side chatter and gracefully resume conversations

2,193 Likes

Cursor adds GPT-5-Codex, a fine-tuned GPT-5 model for agentic coding tasks

4,698 Likes

Google Colab improves notebooks by allowing static runtime versions for consistent results

749 Likes

Reads

Google reports 90% of developers now use AI daily

1,069 Likes

Google’s 2025 DORA report shows how AI transforms delivery: improving speed and reliability, but requiring explicit organizational capabilities. Without these, teams risk stalled adoption and missed opportunities from everyday AI use.

How far can coding agents go improving themselves at runtime?

834 Likes

You learn to build an LLM-powered invoice app, integrate Box MCP server tools for file listing and text extraction, and evolve the system into a coordinated multi-agent design for secure, scalable document workflows.

Google Cloud details tools to build and scale AI agents

1,120 Likes

This talk explores advanced context engineering for AI coding agents. It details strategies like spec-first development, frequent intentional compaction, and subagent workflows. It shows how context management transforms unreliable prototypes into scalable systems for real-world software projects and large codebases.

How To

Build Code from Figma Designs with GPT-5-Codex in Cursor

alpha_signal_image_4

This guide shows how to connect the Figma MCP server and use GPT-5-Codex to turn your designs into code.


Step 1: Install the Figma MCP Server

Go to the Figma MCP Catalog. Follow the setup to add the Figma MCP server to your environment. Add your Figma API key as an environment variable.

Update your servers.json:


{
  "servers": {
    "figma":
  }
}

Step 2: Connect to Your Design

Use the MCP server to fetch a Figma file by URL or file ID. The server exposes frames, layers, and component metadata as structured context.



/call figma.getFile --fileId YOUR_FILE_ID

Step 3: Generate Code with GPT-5-Codex

Prompt GPT-5-Codex via Cursor with the Figma output to generate React, Tailwind, or Next.js code.


Watch now

At Alpha Signal, our mission is to build a sharp, engaged community focused on AI, machine learning, and cutting-edge language models, helping over 200,000 developers stay informed and ahead. We’re passionate about curating the best in AI, from top research and trending technical blogs to expert insights and tailored job opportunities. We keep you connected to the breakthroughs and discussions that matter, so you can stay in the loop without endless searching. We also work closely with partners who value the future of AI, including employers and advertisers who want to reach an audience as passionate about AI as we are.


Our partnerships are based on shared values of ethics, responsibility, and a commitment to building a better world through technology.Privacy is a priority at Alpha Signal. Our Privacy Policy clearly explains how we collect, store, and use your personal and non-personal information. By using our website, you accept these terms, which you can review on our website. This policy applies across all Alpha Signal pages, outlining your rights and how to contact us if you want to adjust the use of your information. We’re based in the United States. By using our site, you agree to be governed by U.S. laws.

Looking to promote your company, product, service, or event to 250,000+ AI developers?

Work with us