Google Jules logo

Google Jules

coding

by google

Async AI coding agent that handles tasks you'd rather not do, powered by Gemini.

Key features

Async Task Execution
GitHub Integration
Critic Agent
Planning Critic
CI Fixer
Suggested Tasks
Pricing

Free tier available, Pro (Google AI Pro) at $19.99/mo

Best For

Developers who want to delegate repetitive coding tasks like bug fixes, dependency upgrades, test writing, and boilerplate generation

Verdict

Free tier with 15 tasks per day and 3 concurrent tasks is genuinely useful for individual developers without any subscription commitment.

What it does

Async Task Execution

Submit a coding task and Jules works in the background. No need to watch or interact. Come back when the PR is ready for review.

Learn more

GitHub Integration

Label issues with 'jules' to trigger tasks automatically. Jules responds to PR comments, fixes CI failures, and creates pull requests with detailed descriptions.

Learn more

Critic Agent

An adversarial code reviewer that runs automatically after Jules writes code. The critic catches bugs, logic errors, security issues, and style violations before you ever see the PR.

Planning Critic

Before executing, Jules generates a plan. The planning critic reviews the plan for feasibility, completeness, and potential issues, catching strategic errors before any code is written.

Learn more

CI Fixer

When a CI pipeline fails on a Jules PR, Jules automatically diagnoses the failure, writes a fix, and pushes a new commit. No manual intervention required.

Learn more

Suggested Tasks

Jules proactively scans your repository and suggests improvements: dependency upgrades, code quality fixes, documentation gaps, and potential bug fixes. Review and approve suggestions to create tasks.

Learn more

Scheduled Tasks

Set up recurring tasks that Jules executes on a schedule. Useful for regular dependency updates, security scans, and automated maintenance routines.

Learn more

CLI and REST API

Interact with Jules programmatically through a dedicated CLI tool or REST API. Create tasks, check status, and integrate Jules into custom automation pipelines.

Learn more

Memory System

Jules remembers your preferences, coding conventions, and feedback across tasks. The memory system builds a persistent profile of how you like code written, reducing the need to repeat instructions.

Environment Snapshots

Save and reuse VM configurations so Jules does not need to reinstall dependencies from scratch on every task. Snapshots speed up task execution for projects with complex build environments.

Built in Browser (Playwright)

Jules has access to a browser inside its VM for testing web applications, verifying UI changes, and debugging visual issues as part of the task execution loop.

Multi Language Support

Works with JavaScript, TypeScript, Python, Go, Java, Rust, and many other languages. Jules adapts to your project's language, frameworks, and conventions automatically.

AGENTS.md Configuration

Place an AGENTS.md file in your repository root to teach Jules your conventions, setup commands, test requirements, and coding standards. Jules reads this file before starting any task.

Learn more

Pricing

Free

Free

Get started with Jules at no cost. 15 tasks per day with up to 3 running concurrently. Powered by Gemini 3 Flash.

  • 15 tasks per day
  • 3 concurrent tasks
  • Powered by Gemini 3 Flash
  • GitHub integration (issues, PRs, comments)
  • Critic agent and planning critic
  • CI fixer
  • AGENTS.md support
  • Memory system
Best Value

Pro (Google AI Pro)

$19.99/month

Higher task limits and concurrency for active developers. Powered by Gemini 3 Pro for stronger reasoning on complex tasks.

  • 100 tasks per day
  • 15 concurrent tasks
  • Powered by Gemini 3 Pro
  • Everything in Free
  • Suggested tasks (proactive scanning)
  • Scheduled tasks
  • Environment snapshots
  • CLI and REST API access
Most popular

Ultra (Google AI Ultra)

$249.99/month

Maximum capacity for power users and teams. 300 tasks per day with 60 concurrent. Currently 50% off at $124.99/mo.

  • 300 tasks per day
  • 60 concurrent tasks
  • Powered by Gemini 3 Pro
  • Everything in Pro
  • Highest concurrency limits
  • Priority task execution
  • Currently 50% off ($124.99/mo)

Pros & Cons

Pros

  • Free tier with 15 tasks per day and 3 concurrent tasks is genuinely useful for individual developers without any subscription commitment.
  • Truly async execution. Submit a task and walk away. Jules works in the background and delivers a ready to review pull request.
  • Critic agent catches bugs, logic errors, and style violations automatically before the PR reaches you, reducing review burden.
  • GitHub native integration: label issues with 'jules' to create tasks, and Jules responds to PR comments for iterative feedback.
  • Fresh VM per task ensures complete isolation. No risk of one task corrupting another or leaving behind stale state.
  • Proactive features (suggested tasks, scheduled tasks, CI fixer) go beyond reactive task execution to actively maintain your codebase.
  • CLI and REST API enable programmatic integration, making Jules composable with custom CI/CD pipelines and automation tools.

Cons

  • GitHub only. Jules does not support GitLab, Bitbucket, Azure DevOps, or other version control platforms at this time.
  • Async execution means no real time collaboration. You cannot pair program with Jules or guide it interactively like you can with terminal or IDE based agents.
  • Paid tiers require a full Google AI subscription (AI Pro at $19.99/mo or AI Ultra at $249.99/mo), not a standalone Jules subscription.
  • The VM environment may not perfectly match your local development setup, potentially causing differences in behavior between Jules's output and your local environment.
  • Enterprise and Workspace support is still in progress. Large organizations with complex compliance and access control requirements may need to wait.
  • No local execution option. Unlike Claude Code or Codex CLI, Jules always runs in the cloud. You cannot run it on your own machine.

How to get started

1

Visit Jules and sign in

Navigate to jules.google and sign in with your Google account. The free tier activates immediately with 15 tasks per day.

2

Connect your GitHub repository

Authorize Jules to access your GitHub repositories. Grant access to the specific repos you want Jules to work with.

3

Submit your first task

Describe a task in natural language: "Fix the failing test in src/utils/parser.test.ts" or "Upgrade all npm dependencies to their latest major versions." Be specific about what you want done.

4

Let Jules work asynchronously

Jules spins up a fresh VM, clones your repo, reads AGENTS.md (if present), plans the approach, executes the task, runs the critic agent, and creates a pull request. You can close the browser and come back later.

5

Review the pull request

When Jules finishes, review the pull request on GitHub. The PR includes a description of changes, the reasoning behind decisions, and test results. Leave comments for adjustments or merge directly.

6

(Optional) Set up AGENTS.md

Create an AGENTS.md file in your repo root with setup commands, test instructions, and coding conventions. Jules reads this file before every task, ensuring consistent behavior.

7

(Optional) Try the CLI or REST API

For programmatic access, install the Jules CLI or use the REST API to create tasks, check status, and integrate Jules into your automation workflows.

Deep dive

Detailed guides with comparisons, tips, and visuals for each feature.

Get notified about updates

We'll email you when this tool's pricing or features change.

Last updated: 2026-02-21