Jules
Async coding that opens PRsGoogle's asynchronous coding agent that clones a GitHub repository to a cloud VM, drafts a plan, and opens a pull request for review.
- Maker
- Google (Google Labs)
- Founded
- 2024
- Pricing
- Free tier
- Last checked
- October 2026
Verdict
Google's asynchronous coding agent that clones a GitHub repository to a cloud VM, drafts a plan, and opens a pull request for review.
Score breakdown
vs Claude Code, the category leaderDevelopers who want to hand well-defined coding tasks to an agent that works in the background and returns a pull request
Key features
- Asynchronous task execution on a cloud VM cloned from your GitHub repository
- Plan-then-approve workflow: review the proposed plan and diff before changes land
- Creates pull requests you can review, approve and merge on GitHub
- Assign tasks from a GitHub issue using a label
- Parallel tasks, with concurrency limits set by plan
- Gemini model access, with higher-tier plans getting newer models first
Strengths
- +Free plan with 15 tasks per day and 3 concurrent tasks, so it can be tried without payment
- +Works asynchronously in a cloud VM, so you can assign a task and review the result later
- +Produces a reviewable plan and diff before any changes are approved, then opens a pull request on GitHub
- +Bundled with Google AI Pro and Ultra subscriptions, so one plan covers Gemini and other Google tools
Weaknesses
- −Limited to GitHub repositories, so teams on other source hosts cannot use it
- −Daily task caps on lower tiers can run out quickly on busy days
- −Result quality depends on how well the task is specified, and failed runs need a rewritten prompt
- −Model names and tier details changed quickly in 2026, and third-party sources disagree on some of them
Overview
Overview
Jules is an autonomous coding agent from Google Labs. You point it at a GitHub repository and a branch, write a task (or label a GitHub issue), and it clones the repository to a cloud virtual machine. It then drafts a plan, makes the changes, and proposes a diff. Once you approve it, Jules opens a pull request that you can review and merge on GitHub.
Its main difference from interactive coding tools is that it runs asynchronously. You do not need to watch it work, which makes it a fit for well-scoped tasks such as tests, small bug fixes and dependency updates. Access is tiered by daily task count and concurrency, and the free tier allows 15 tasks per day. The official plan page does not list prices, so the monthly prices above come from third-party listings and should be checked against Google’s current plan pages.
Key Benefits
- Background execution: Tasks run in a cloud VM, so developers can keep working while the agent runs.
- Reviewable output: Each task ends in a pull request, so changes go through the same review process as human code.
- Low entry cost: The free tier allows real use without a subscription.
Use Cases
- Routine maintenance: Assign test additions or small refactors and review the resulting pull request.
- Issue-driven work: Label a GitHub issue to have Jules attempt a fix and open a pull request.
- Parallel tasks: Run several well-defined tasks at once within the concurrency limit of your plan.
Pricing
Checked October 2026 · USD- 15 tasks per day
- 3 concurrent tasks
- 100 tasks per day
- 15 concurrent tasks
- Higher access to the latest Gemini models
- 300 tasks per day
- 60 concurrent tasks
- Priority access to the latest Gemini models