Refact.ai Changelog: How to Simplify Your Workflow and Take Control of Your Data
October 29, 2024
by Katrin Maikova
4 min read
TL;DR
New Diffs Workflow: Ask Refact.ai to handle tasks for you and quickly spot code changes with the new Diffs workflow.
Code Lens Commands: Fix bugs, get code explanations, and customize your workflow—all in one click.
New Privacy Settings: Take full control of your data with customizable privacy rules.
First AMA on Discord: Join us to see all the new features in action.
Let’s take a closer look!
🔥 Master Code Changes with New Diffs Workflow
Ask Refact.ai to manage tasks for you and quickly spot code changes using the newly launched Diffs workflow. This workflow provides a clear, visual comparison of code modifications, enabling faster and more efficient development.
It simplifies change management and provides clear visibility into updates for tasks such as:
Refactoring code
Write docstrings
Writing tests
Rewriting code in a different language
Any other task you can express through a natural language request!
Here’s something even better: you can attach a file for Refact.ai to process or ask it to scan the entire repository for specific tasks.
With the Diffs workflow, your development process becomes faster and more reliable, ensuring clear visibility into every update.
🔎 Customize Code Lens Commands
Now your coding experience is even more intuitive with our Code Lens Commands. As you write code, you can instantly call a chat to find bugs or ask for an explanation of any code snippet.
That’s not all—you can also customize it to handle your most frequent tasks, such as creating unit tests in your team’s preferred style or generating test plans.
Simply navigate to Settings -> Edit customization.yaml and configure your own custom buttons.
🗝 Take Control of Your Privacy
Your privacy is a priority for us, and we’re giving you full control over it.
You can specify rules by going to Settings -> Edit privacy.yaml to prevent certain files from being sent. Files marked as ‘blocked’ will never leave your device; however, they won’t be used to improve the quality of chat responses, and code completion will not be available for them.
If you want some files to only be sent to servers you control (e.g., self-hosted with your own key), use the only_send_to_servers_I_control setting. Code completion will still work for these files, but they won’t be used for chat with GPT or other third-party models.
For users who prioritize privacy, we offer a self-hosted version of Refact.ai. By deploying it on your own server, all your data remains fully under your control.
👀 Watch Refact.ai in Action at the AMA (Ask Me Anything) Event
At Refact.ai, our goal is not just to empower developers with AI but to build a vibrant community of like-minded developers. We’re uniting those who are passionate about building the future of programming and pushing the boundaries of what we can achieve together with AI.
Get expert support, share knowledge with peers, and learn about updates firsthand. Ready to dive in?
Code completions powered by the StarCoder2-3B model.
Access to the best models including GPT-4o, GPT-4 Turbo, GPT-4o-mini, Claude 3.5 Sonnet, Groq-Llama-3.1
4096-token context length for more accurate code completions (compared to 2048 tokens in the Free plan).
32k-token context length for enhanced chat interactions (compared to 8k tokens in the Free plan).
Codebase-aware AI for both completions and chat (RAG).
All of this leads to more accurate auto-completions and higher-quality chat responses. Use promo code OCTOBER to enjoy your first month free, and afterward, it’s just $10/month for all these powerful features!