Jules 3.0 + Jules Tools CLI: This with Gemini 3.0 WILL BE A BANGER!

Key Concepts

• Jules: An AI coding agent, praised for its UI and experience, used for remote coding tasks.
• Gemini 2.0/2.5 Pro: AI models mentioned in the context of Jules's initial teasing and capabilities.
• Pull Request (PR) Integration: The ability for Jules to interact with and fix issues directly within GitHub pull requests.
• File Selector: A feature allowing users to precisely specify which files Jules should work on for a given task.
• Memory Feature: Jules's ability to learn from user interactions (preferences, nudges, corrections) and apply this knowledge to similar future tasks within the same repository.
• Jules Command Line Interface (CLI) / Jules Tools: A new interface for direct control over Jules, enabling scripting, customization, and integration into existing workflows.
• Patch Application: The CLI's capability to instantly pull work-in-progress code from an active Jules session and apply it locally for immediate testing.
• Scriptable and Composable: The CLI's design allowing integration with other command-line tools (e.g., GH, JQ, CAD).
• Terminal User Interface (TUI): An interactive dashboard within the CLI for guided task creation and management.
• Jules API: An application programming interface for integrating Jules into other systems.
• Ninja Chat: An all-in-one AI platform offering access to various AI models and tools.

Jules: An Advanced AI Coding Agent and Recent Upgrades

The video discusses recent significant upgrades to Jules, an AI coding agent, highlighting its capabilities and user experience. The speaker considers Jules to be one of the best remote coding agents available, even surpassing others like Codeex, particularly in terms of UI and overall experience, setting aside raw model capability. These updates were part of a recent "Jules week."

I. Enhanced Integration and Control

• Pull Request (PR) Integration: Jules can now be integrated into pull requests. Users can delegate issues to Jules within a PR branch, and Jules will fix them while updating the pull request accordingly.
  • Default Behavior: By default, Jules registers for all comments on any pull request. If a change is requested in a comment, Jules will attempt to implement it.
  • Customization: This behavior can be modified in settings to make Jules respond only to comments that explicitly mention @Jules.
  • Note on @Jules Tag: The speaker points out a minor UI anomaly where clicking the @Jules tag in GitHub navigates to a specific GitHub user ID, which feels "very odd" and needs a fix.
• Granular File Selection: A new file selector allows users to tell Jules exactly which files to work on for any given task. This feature removes ambiguity, provides more granular control over Jules's actions, and helps tighten the context for specific changes.
  • Usage: Users can hit the "attach" option in the Jules prompt box and select an exact file or folder.

II. Memory Feature for Personalized Learning

Jules now includes a "memory" feature designed to learn from user interactions.

• How it Works: During a task, Jules saves user preferences, nudges, and corrections.
• Benefits: The next time a user runs the same or a similar task in that specific repository, Jules references its stored memory to better anticipate needs and follow established patterns, leading to more accurate results with less guidance.
• Management:
  • Toggle: Memory can be toggled on or off for a repository via the repo settings page under the "knowledge" tab.
  • Viewing and Editing: Users can view all memories created by Jules by navigating to their codebases, then the "knowledge" tab. From there, memories can be removed, new ones created, or existing ones slightly edited.
• Analogy: This feature is compared to how "byte rover" or "super memory" agents operate, where the agent performs a tool call to save memory when needed and references it later. The speaker notes it's "pretty great" but can be "somewhat finicky" if not implemented correctly.

III. Jules Command Line Interface (CLI) / Jules Tools

A major update is the launch of Jules Tools, a new command-line interface (CLI) designed to provide direct control over the AI coding agent.

• Purpose: The CLI makes Jules scriptable, customizable, and easy to integrate into existing workflows.
• Key Functionalities:
  • Direct Control & Task Creation: Users can create tasks directly from the command line.
  • Session Management: Commands like Jules remote new and Jules remote list allow users to list and monitor active Jules sessions without leaving the command line.
  • Local Patch Application: Users can instantly pull work-in-progress code from an active Jules session and apply it to their local machine. This enables immediate testing of changes without waiting for a commit to GitHub.
  • Scriptable and Composable: Jules can be integrated into automations by piping in output from other command-line tools such as GH (GitHub CLI), JQ (JSON processor), or CAD.
  • Interactive Dashboard (TUI): For a more guided experience, the CLI includes a built-in Terminal User Interface (TUI) to create and manage tasks step-by-step.
• Installation: The CLI can be installed easily with a simple command.
• Practical CLI Usage:
  • Jules command: Running Jules displays all previously completed tasks along with their status.
  • Task Interaction: Users can click on a task to open its page in the browser or easily clone the code with changes to their local machine.
  • /new command: This command allows users to create a new task directly from the CLI and delegate it to the online coding agent.

IV. Jules API

Jules has also launched an API, making it very easy to integrate into almost any system. The speaker notes its simplicity and ease of use, reinforcing the idea that Jules feels like a "proper product."

V. Ninja Chat: An All-in-One AI Platform (Sponsor Segment)

The video includes a sponsored segment for Ninja Chat, an AI platform offering access to various AI models and tools.

• Features:
  • Access to top AI models like GPT-4o, Claude for Sonnet, and Gemini 2.5 Pro.
  • AI playground for side-by-side comparison of responses from different models.
  • Mind map generator for organizing complex ideas.
• Pricing:
  • Basic plan: $11 per month, includes 1,000 messages, 30 images, and 5 videos monthly. Higher tiers are available.
• Discount Codes:
  • king25: 25% off any plan.
  • king40 yearly: 40% off annual subscriptions.

VI. Conclusion

The speaker expresses strong positive sentiment towards Jules's recent updates, particularly the CLI and API, which transform it into a more robust and integrated product compared to alternatives like Codeex. The updates provide users with greater control, efficiency, and a more personalized experience, making Jules a powerful tool for remote coding.