Today we’re excited to launch Kite Team Server, the industry’s first self-hosted ML engine for 4 token autocomplete technology.
TL;DR
- Kite Team Server is a self-hosted version of Kite for the enterprise that provides personalized code completions in 16 programming languages across 16 editors/IDEs.
- Kite Team Server uses the latest developments in machine learning to “custom-train” ML models for each company, providing personalized completions based on a company’s codebase.
- The experience builds on the existing UI of Kite Free, which 400,000 developers use to code 18% faster each month, augmented with enterprise-grade security and completions tailored to your company’s coding style.
Support for your entire team.
Languages
- Python
- JavaScript
- Java
- HTML / CSS
- C Based
- Go
- TypeScript
- Scala
- Kotlin
- PHP
- Ruby
- Bash
Editors
- Jupyter Lab
- VS Code
- IntelliJ
- PyCharm
- Sublime
- Spyder
- WebStorm
- Vim
- Atom
- CLion
- PHPStorm
- Rider
- RubyMine
- AppCode
- GoLand
Kite Pro: ML Model on Your Laptop
Kite Team Server: 25X Larger ML Model in the Cloud
Kite Team Server Without Custom Model Training
Kite Team Server With Custom Model Training
- Larger ML Models
- Personalized Completions
The free version of Kite is trusted by over 400,000 developers, including software engineers at 35% of the Fortune 500. A key feature request from our users has been a version of Kite built for the enterprise, helping teams use Kite to code faster. By reimagining the Kite completions engine for the enterprise, we built a solution that is nimble enough to provide value to teams of just a few developers and robust enough to support thousands of simultaneous users. Our new enterprise-grade application architecture satisfies the most rigorous IT and security requirements.
Kite Team Server provides better completions, more often. Underpinning this is an advanced GPU that sits inside your company’s internal network. Kite Team Server uses this GPU to train personalized ML models on a company’s code base 1The automated training process completes in one hour and requires no intervention or ML background from the team admin.. Completions shown by Kite Team Server guide developers to repeat idioms and patterns from internal code. The GPU also uses a server-scale machine learning model to process all completions requests 2In contrast, Kite Free and other completions engines use the CPU on user’s computers to process completions. Since completions processing is offloaded to the server, Kite Team Server uses less battery, CPU, and memory than any other completions engine. . Compared to our offering for individuals, the average completion length doubles from two words to four words and completions used per line of code written increases by 40%. These code suggestions help developers stay in flow and write code faster.
Kite Team Server provides completions in all of the most popular programming languages and IDEs. Our users don’t write code in one language, or even one editor. Teams that use Kite can leverage our completions engine across a wide range of tasks, thanks to built-in support for 16 languages and 16 IDEs (even Vim!).
We can’t wait to bring Kite Team Server to your company. Get in touch or watch the webinar.