Kilo Code
Open-source AI coding agent
Kilo Code is an open-source artificial intelligence coding agent for software development. It can generate and modify source code, execute commands, debug software, and perform development tasks from natural language instructions. Kilo Code is model-agnostic and can work with hundreds of large language models rather than being tied to a single AI provider.
The software is available as an extension for Visual Studio Code and JetBrains integrated development environments, through a command-line interface, and through cloud-based agents. Kilo Code was acquired by Anaconda in July 2026.
01History
Kilo Code was launched in March 2025. It began as a fork of the open-source coding agent Roo Code, which was itself derived from Cline. Kilo subsequently incorporated features from both projects while developing its own agent and model-routing tools.
The company was founded by Scott Breitenother, Emilie Schario, and Sid Sijbrandij, a co-founder of GitLab. In December 2025, Kilo Code raised US$8 million in seed funding led by Cota Capital, with participation from General Catalyst and other investors.
On July 15, 2026, Anaconda acquired Kilo Code. At the time of the acquisition, the software was reported to be used by more than three million developers.
02Features
Kilo Code can inspect and edit files in a software project with an agentic harness, generate and refactor code, execute terminal commands, debug programs, and automate browser interactions. It provides specialized modes for tasks such as planning, coding, and debugging, and can delegate parts of a larger task to subagents.
The software can be used in Visual Studio Code, JetBrains IDEs, and from a command-line interface terminal. Kilo also provides cloud agents that can perform development tasks remotely.
Kilo Code also provides a mobile app that can connect to cloud agents and remote sessions running through its CLI or editor extensions, allowing users to monitor and continue coding sessions from a mobile device.
AI models and providers
| Supported external providers and plans | |
| Standard API keys | Subscription and direct provider plans |
|---|---|
|
|
Kilo Code separates the coding agent from the large language model used to perform a task. Its Kilo Gateway provides access to hundreds of models, while users can alternatively connect through OpenRouter, connect up to 30 pre-existing subscription plans from AI providers, and an automatic model-selection feature Auto Model can route coding tasks to different models based on factors such as the type of task, performance, and cost.
Kilo Code can also use locally hosted models through Ollama and LM Studio, or connect to OpenAI-compatible local inference servers such as vLLM and SGLang.
Sources and credits
This article is adapted from the Wikipedia article “Kilo Code”, written by its contributors and licensed under CC BY-SA 4.0. Fathomly has changed the layout, removed citation markers, navigation and maintenance notices, and adjusted punctuation. This adapted version is shared under the same license. For references, see the original article.
Images, from Wikimedia Commons:
- Kilo Code CLI.png by Kilo Code GitHub repository, MIT
Fathomly is not affiliated with or endorsed by the Wikimedia Foundation. Spotted a problem? Tell us.
