Refact.ai Updates for June 2024

It’s time for our monthly roundup of the latest improvements and features the Refact.ai team has been working on! In June we added Claude 3.5 Sonnet, upgraded our JetBrains chat, expanded some models’ context, and more. Let’s dive into what’s new!

New Model: Claude 3.5 Sonnet

We’ve added support for Claude 3.5 Sonnet in Refact.ai’s chat—the latest and most powerful AI model from Anthropic! It outperforms Claude 3 Opus and GPT-4o in nearly all benchmarks.

With Refact.ai chat, Claude 3.5 can leverage your codebase context for even faster and more accurate coding assistance. We’ve tested it, and it’s truly impressive!

Our in-IDE chat now offers these AI model options:

• GPT-3.5
• GPT-4 Turbo
• GPT-4o
• Claude 3.5 Sonnet

Easily switch between them for various tasks like asking questions, explaining code, debugging, generating new code, creating documentation, and writing unit tests.

Claude 3.5 Sonnet is available to users of our Refact.ai Pro, Team, and Enterprise plans. Try it for free with a promo code on our Plans page!

Improved Chat for JetBrains Plugin

Our JetBrains in-IDE chat got a complete makeover! It’s now more powerful and understands your entire codebase for contextual answers.

New features include:

• @-commands: Easily add your codebase or specific files as context in chat. Learn more here
• Custom prompts: Create chat prompts tailored to your unique coding needs.

This update marks a significant milestone: Refact.ai’s in-IDE chat for JetBrains now matches the functionality we offer for VSCode—much to our delight as JetBrains fans ourselves!

More Plugins Updates

• AST File Limit (JetBrains)
  A new setting for optimized performance with large codebases.
• Ast-light mode (VS Code)
  Working with a massive repo or on a machine with limited RAM? Turn on the new ast-light mode in Settings for efficient memory usage.

Enterprise and Self-hosted Updates

Docker v1.6.0:

• 8k context support for Llama3 models
• Enhanced token and key information in Settings → Credentials

For Enterprise users:

• 4k context support for Refact vllm model
• Improved integration security when creating an AWS image

Coming Soon: Pre-release Features

Get an early look at our upcoming features! You can try these out now by switching your Refact.ai extension to the Pre-release version:

• Chat functionalities hints
• ’Use tools’ in chat: Enable AI model to work autonomously with your codebase context. It can perform multi-step operations, deciding what data it needs for your request, finding it, and using it—all on its own.

Refact.ai: An Open-source Copilot Alternative with Customization and Advanced AI Coding Features

That’s the scoop for June—hope you like our updates!

If you want to try Refact.ai today, visit our Pricing page for 1 month free. Got a whole team of developers? We’d be happy to show you how Refact.ai can help your software company boost performance with AI. Just set up a demo call with our team.

We’d love your feedback:

• Join our Discord to discuss new functions or share feedback
• Check Refact.ai GitHub for our source code — Refact.ai is an open-source AI coding assistant and welcomes your contributions!
• Please rate us on VS Code and JetBrains
• And stay (fine-)tuned for further updates!