Overview
Jules is Google’s autonomous coding agent, built around a different philosophy than editor assistants like Cursor or Copilot. Instead of sitting in your editor, it works in the background: you point it at a GitHub repository, describe a task in plain language, and walk away. Jules clones the repo into an isolated Google Cloud VM, writes a plan for your approval, makes the changes, runs a built-in critic pass and your test suite, then opens a pull request. Since 2026 it also fixes its own failing CI runs and resubmits. It graduated from Google Labs to general availability at Google I/O 2026, powered by Gemini 3.1 Pro on paid tiers.
Key Features
- Async cloud execution — Jules clones your repo into a sandboxed Google Cloud VM and works off your local machine, so you can close the laptop and come back to a finished PR.
- Plan-then-execute — before writing code it shows a step-by-step plan you can review and steer, reducing surprise changes.
- Built-in Critic + test runs — a reviewer agent checks changes and Jules runs your existing suite, catching regressions before the PR lands.
- CI auto-fix — on its own pull requests Jules detects and repairs failing CI, then resubmits, closing the loop most async agents leave open.
- GitHub-native — start a task by labeling an issue with ‘jules’; Jules opens branches and descriptive PRs that read like human contributions.
- CLI, REST API (alpha) and MCP — extend Jules beyond the web UI into Slack, CI/CD, and vetted partner tools like Linear, Neon, and Supabase.
Pricing
| Plan | Price | For |
|---|---|---|
| Free | $0 | 15 tasks/day, 3 concurrent, Gemini 3 Flash base model |
| Pro (Google AI Pro) | $19.99/mo | 100 tasks/day, 15 concurrent, Gemini 3.1 Pro |
| Ultra (Google AI Ultra) | $99.99/mo | 300 tasks/day, 60 concurrent, priority model access |
Comparison
vs. Claude Code: Claude Code lives inside your terminal for synchronous, step-by-step control — great when you want to steer every edit. Jules is the opposite: asynchronous and GitHub-native, better for fire-and-forget backlog work than tight pair programming.
vs. Cursor: Cursor embeds AI completions directly in your editor for real-time help. Jules doesn’t sit beside your cursor; it works in its own VM and returns a PR, so it suits parallel maintenance rather than live coding.
vs. GitHub Copilot: Copilot suggests code as you type inside the IDE. Jules takes a whole scoped task off your plate and delivers a reviewed, tested pull request — complementary rather than competing for the same moment.