Cascade on JetBrains Launch 🚀

Introducing the flow evolution of Chat: Cascade

Cascade's agentic capabilities unlock a new level of collaboration between AI and human, making it the ultimate partner for complex coding workflows.

Write & Chat Mode

• Write mode: Enables Cascade to create and modify codebases.
• Chat mode: Optimized for codebase questions and general coding principles.

Model Selection

• Added support for new models including Claude 3.7 Thinking, GPT-4.5, Deepseek V3 and more.

Access to Tools

• Added multiple tool integrations including Search, Analyze, Web Search, and Terminal access
• Added image upload support for premium models.

Revert to previous steps

• Ability to revert edits that Cascade has created at any point in the conversation.
• Hover over the original message and click on the revert arrow on the right or revert from the table of contents.

Contextual Awareness

• Contextual awareness capabilities that track IDE state changes in real-time.
• Reduced need for explicit context prompting due to improved awareness.

Teams Admin Controls

• For teams and enterprise plans, admins can select which models are available to its team.
• As a member of a team, you will only see the models that are available to your team unless enabled by your admin.
• Admins can find these controls at https://codeium.com/team/windsurf_settings under "Extension Models".

Usage Transparency

• Added credit usage display for Cascade actions
• Updated usage pricing system for Windsurf Plugin. See pricing for details
• Credit usage can be monitored via the Plan Info tab in the Windsurf widget on the bottom toolbar.

Codeium is Now Windsurf

• We are renaming our company to Windsurf & our extension product to Windsurf Plugin.
• Since the launch of our Windsurf Editor, we have captured what we are really building: combining human ingenuity and machine to result in experiences that feel and appear effortlessly powerful.

What's New with Codeium v1.38 🚀

Bug Fixes

• Performance improvement on large repositories.
• Indexing Performance Boosts.

What's New with Codeium v1.36 🚀

Bug Fixes

• Performance improvement on large repositories.
• Improved stability across network disconnections.
• Fixed changelog redirect page.
• Improved extension logging.

What's New with Codeium v1.30 🚀

JetBrains Pre-Release 🪄

We're excited to introduce the Pre-Release Channel for JetBrains! Get early access to upcoming features and help shape the future of Codeium.

• Switch between stable and pre-release versions
• Test new features before they're officially released.
• Provide valuable feedback for improvements.

Bug Fixes

• Disabled Codeium functionality in Python Debugger Console.
• Fixed Download Diagnostic button in chat client.

JetBrains-specific updates

• Added Codeium action group to Tools menu
• Added "Codeium"/"Codeium Teams" at bottom right corner
• Disabled inline shortcut hint when conflicting with vim plugin
• Added refresh chat client action to Tools menu
• Added manual autocomplete trigger
• Improved recognition for IDE-unrecognized languages
• Add re-signin instructions when API key is invalid

Enterprise & technical updates

• Added enterprise Grafana dashboard panels for index service tracking
• Fixed project closure/reopening bug in JetBrains command
• Converted reads from subscriptions to users
• Added dev mode and teams mode to unleash in VSCode
• Fixed analytics API typo and PCW calculation error
• OCaml interface and OCaml files now deemed relevant to each other
• Allow specifying chat-related ports in LS
• Emacs and Vim extensions on enterprise no longer need specific tag checkout
• Vim extension in enterprise-mode automatically downloads enterprise language server
• Automatic deletion of corrupt SQLite databases in language server
• Fixed batch deletion button in indexing UI
• Enhanced telemetry batching
• New model image handling and inference server updates for Kubernetes deployments

Codeium Command UX Upgrade

Codeium Command got a major renovation.

The most salient features are:

• Persistent experience: Opening Command now instantiates a popup that lives near the relevant code until you dismiss it, instead of dissipating when you click away.
• Follow-up support: You can now append instructions to your initial one and rerun the new command on the original code.
• Hotkeys: Escaping, submitting commands, undoing or canceling commands, toggling focus between the popup and the editor, and accepting commands can all be carried out from the keyboard.
• Block undo: Ctrl+z or ⌘+z while in the editor after a Command output will undo Command's entire change. You can also Ctrl⌫ or ⌘⌫ for the same effect while rehighlighting the original text and popping the latest instruction so that it's ready for a different command.

Chat Context Upgrade

Our latest update supercharges your chat experience with a suite of new actions:

• Copy to clipboard: Easily copy messages with a single click.
• Rerun with context: Rerun messages including relevant code from your project for improved responses.
• Retry Button: Quickly retry messages that encountered errors.
• Stats for nerds: Peek behind the scenes with detailed stats on what's happening under the hood.
• /explain command: Get detailed explanations for functions, classes, and parts of your code right within the chat.
• @mention files and directories: Easily reference files and directories in your code.
• Inline context item pinning: Pin important context items for quick access.

We've also engineered the Chat input text editor to be significantly faster and more responsive.

Nudge

Highlighting a section of code now conveniently opens a small popup tool that lets you directly ask about the code in Chat or run a Command on it. The nudge can be disabled in your Codeium Settings.

Codeium Command

Command gives developers the ability to instruct the AI to perform tasks within the editor. Just open the command prompt (Ctrl+I or ⌘+I for Mac), enter a command, and watch Codeium code for you!

And even more powerful is providing in-line instructions on existing blocks of code in order to perform direct transformations. A simple one would be to just add comments to your code:

Guidelines For Chat

Enhance your chat experience with custom context! In the chat panel, you now have the option to provide your own short snippets of context for the AI.