Refact.ai Huge Updates: JetBrains Release, Bring Your Own API Key, Upcoming Agentic Features
TL;DR
- Largest JetBrains Update Ever: Completely new user-friendly interface, all the functions previously available to VSCode users, are now available for JetBrains users.
- Better UI, Onboarding, and Tip of the Day: It looks and works better, the new step-by-step tour, improved user interface, and daily productivity tips.
- Bring Your Own Key: Use your existing API key with Refact.ai to use almost any server or service for inference.
- Try the Agent’s Capabilities First: Explore the AI Agent features and give us your feedback.
JetBrains Users: The Biggest Update Ever is Here!
We value each of the over 37,000 people who have trusted us and downloaded Refact for JetBrains, and we have exciting news for all of you! The user interface is now better than ever, with features previously accessible only in the VS Code environment now available to all JetBrains users!
Boost Productivity with Refact.ai: Improved UI and Step-by-Step Onboarding
Want to easily discover Refact.ai’s main capabilities?
Now, when you start using Refact, we’ll guide you through all the product features designed to help you solve tasks more effectively.
All the capabilities listed below are now available for both VS Code and JetBrains:
1. Accurate Auto-completion
Our auto-completion goes beyond standard coding assistants. Using Retrieval-Augmented Generation (RAG), Refact analyzes your entire codebase to provide more accurate and context-aware suggestions. This approach significantly improves the relevance and precision of coding recommendations.
2. Context-Aware Chat
Imagine you’re deep in a complex project—our context-aware chat, already familiar with your codebase, can pull up relevant code snippets, saving you time and reducing errors. Simply open the chat and ask questions or assign Refact.ai to:
- Find and fix bugs,
- Write documentation,
- Create unit tests,
- Review code,
- Explain highlighted code snippets,
- Summarize code in concise paragraphs,
- And more!
3. Access to the Latest Language Models
- With the Free plan, you have GPT-4o-mini by default.
- The Pro plan gives you access to more models, such as GPT-4o, GPT-4-turbo, and Claude 3.5 Sonnet.
Each model has its strengths, so you can try and choose the one that best fits your needs.
See for yourself—open Refact.ai in your favorite IDE and let our guided tour show you the product’s power.
Your Daily Dev Boost with Tip of the Day
Starting now, we’ll provide you with daily tips to make your experience with Refact.ai even more powerful.
If you have a tip that could improve Refact.ai or an impactful use case, share it in our Discord community to help others get the most out of their experience!
Bring Your Own API Key and Use Refact.ai with Your Favorite Models
Already have an API key? Now you can use the Refact.ai plugin with any models you prefer! Bring your own API key to Refact.ai and use any OpenAI- or HuggingFace-style APIs, including Amazon Bedrock, llama.cpp, OpenRouter, and many more.
Did you know that with the Refact.ai Pro plan, you can access all GPT models and Anthropic’s Claude 3.5 Sonnet for just $10? Using these models separately would cost $20 each—totaling $40. By using Refact.ai, you save $360 per year and gain access to the best models all in one place. Additionally, these models will be fully aware of your codebase, providing more accurate and context-aware assistance tailored to your specific needs.
With promo code MODELS, you can get 1 month for free!
One More Step Toward the Future of Programming
With Refact.ai, you now have access to three distinct modes: Quick, Explore, and Agent. Here’s how they differ:
Quick Mode: In Quick mode, the model responds instantly without accessing external tools. However, you can still enrich your interaction with @-commands, such as:
- @web: Easily convert any webpage into plain text for quick summarization and interaction. Just use @web followed by the URL (e.g., @web https://refact.ai/) to fetch and convert the content into text that can be used within your chat.
- @search: Quickly locate similar code or text within your workspace, directory, or file. Simply use @search followed by your query (e.g., @search create table) to find matching content for seamless exploration and interaction.
Explore mode: More advanced than Quick mode.
- Utilizes exploration tools to gather context about the project before answering questions.
- Automatically uses @-commands, such as search, definition, reference, and tree, to explore the codebase and provide more informed answers.
- Can look into multiple files and understand the project structure.
Agent Mode: The model has agent capabilities, which completely change the way you program. While responses may take longer, it provides higher-quality solutions for your complex challenges. We’re still working on it, but it can already:
- Make code changes based on natural-language prompts directly from the chat,
- Browse the internet to find the best solutions for your tasks,
- And even write prototypes!
We’re just getting started! Stay (fine-)tuned as we continue to build the future of programming.
- Join our Discord to discuss new functions or share feedback
- Get Refact.ai for your dev team: Set up a quick demo call with us and get a 2-week trial of Refact.ai Enterprise.