What is Jules? Google's New Asynchronous AI Coding Assistant That’s Transforming Development

Jules is an AI for coding agent developed by Google and opened for public beta in May 2025. Unlike real-time autocomplete tools, Jules is a "task-oriented, asynchronous" intelligent coding agent.

Jules works by cloning your GitHub project into a secure cloud-based virtual machine (VM), where it executes multi-step development tasks based on your input. Whether refactoring code, upgrading versions, or writing unit tests, Jules automates it all and generates both change reports and audio changelogs to facilitate developer review.

Powered by Google's advanced multimodal large language model, Gemini 2.5 Pro, Jules can reason across files, understand project history, and track logical dependencies—allowing it to comprehend entire projects, not just code snippets.

Image source: Jules

Jules doesn’t just help you “write a line of code.” It’s built to understand project context and autonomously manage complex tasks within secure Google Cloud VMs , with isolated execution for privacy and security. Key features include:

• Multi-step task execution : From upgrading Node.js to writing tests, all in a single task.
• Bug fixing & dependency upgrades : Automatically locates bugs, patches them, and updates dependencies.
• Audio changelogs : Converts code changes into audio summaries for better team communication.
• Full PR generation with diff comparison : Creates pull requests automatically after task completion.
• Deep GitHub integration : Works directly within GitHub without extra installations or switching environments.

Image source: Jules

Jules is an experimental asynchronous AI Coding assistant that helps you fix bugs, write docs, or build features—fully integrated with GitHub . It clones your repo into a VM, handles the job, and lets you work on other tasks while it runs. Here's the full workflow:

1. Log into Jules and connect your GitHub account.
2. Select a repo and branch : Any repo or feature branch is supported.
3. Input task instructions : e.g., “Upgrade to Node.js 18 and fix compatibility issues” or “Add tests for login API.”
4. Review the task plan : The system outlines the files to modify, steps, and expected results.
5. Run the task in background : Jules processes everything in a VM while you continue other work.
6. Review results and merge PR : You’ll be notified upon completion. Review and merge if everything looks good.

You can also set up your environment using a setup script (e.g., npm install, npm run test). All tasks run in short-lived Ubuntu VMs with support for JavaScript, Python, Go, Java, Rust, and more. Use “Run to Validate” to test your setup script before running tasks.

Jules stands out with asynchronous task handling, multi-goal execution, and built-in changelogs for team communication . Here's a comparison chart with GitHub Copilot , OpenAI Codex , and Devin:

Jules shifts the paradigm from AI as a tool to AI as an autonomous agent . Instead of manually directing every step, developers simply define goals and let Jules handle the process—from planning to execution.

This significantly reduces repetitive work, allowing developers to focus on strategy, architecture, and innovation. It's a major leap toward smart development .

Besides task-driven tools like Jules, ChatApe by GenApe enables developers to interact with AI using natural language to complete coding tasks—ideal for those who prefer chat-based development.

Start your AI Coding journey now with ChatApe: https://app.genape.ai/chat-ape

All Jules operations run in isolated VMs provided by Google Cloud. Each task is sandboxed, and no private data is used for AI training . All data is temporary and deleted after use.

Jules operates strictly within the authorized GitHub scope. Before starting any task, it generates a detailed plan outlining intended changes and reasoning. Users can modify or cancel tasks at any time, ensuring final control remains with humans.

Currently in public beta, Jules is free to use globally. Each user can run up to 5 tasks per day, with a maximum of 3 concurrent tasks and 5 codecasts daily .

If limits are exceeded, new tasks are disabled temporarily. Existing tasks can still be managed, and history is preserved.

Google states that current usage data will guide product refinement, performance tuning, and future pricing models . Paid plans are expected based on feedback and usage patterns.

Enterprise users and power developers can apply for increased task limits by submitting a form detailing repo usage and desired capacity.

Early beta users generally praise Jules for saving time on large projects, upgrades, testing, and refactoring. Some note the UI could be improved and complex task planning needs refinement.

Jules represents a new generation of agentic AI for coding . It’s not just a helper—it’s an intelligent agent capable of independently executing complex tasks. For individual developers, it saves time. For teams, it boosts quality and efficiency.

If you're curious about AI-based development, try ChatApe by GenApe for a more intuitive, chat-style experience—no commands needed, just natural conversation.

Start your AI Coding journey with ChatApe now: https://app.genape.ai/chat-ape