Kilo Code bets on agentic engineering with model-agnostic CLI (www.fastforward.blog)

Kilo Code, co-founded by GitLab chair Sid Sijbrandij, has launched a new command-line interface (CLI) designed to work seamlessly with over 500 AI models. This model-agnostic tool enables developers to build and manage agentic workflows, allowing them to choose the best AI model for each specific task. Co-founder Scott Breitenother emphasizes the importance of "model freedom," facilitating a shift from traditional chatbot interfaces to a platform where developers can orchestrate multiple agents simultaneously, akin to conducting an orchestra. The CLI integrates with popular coding environments like VS Code and JetBrains, enhancing development workflows by enabling long-running sessions and parallel agent management. Kilo Code is committed to an open-source ethos, operating under an MIT license and prioritizing transparency in contrast to the "black box" nature of many modern AI solutions. The platform allows developers to customize their workspaces with specialized modes tailored to different tasks, such as Code, Ask, Architect, Debug, and Orchestrator. While Kilo charges for AI model usage at pass-through pricing, the startup aims to monetize through enterprise agreements. With over a million downloads to date and a recently completed $8 million seed round, Kilo Code is resonating well within the developer community as it sets the stage for more flexible and user-controlled AI development.