Kilo Code is Kilo Code Inc.’s open-source AI harness for VS Code, JetBrains and the terminal: a coding agent that makes language models operational through MCP tools, skills and a permission system - with free choice among more than 500 models. As of 20 September 2026.
What is Kilo Code?
Kilo Code is developed by Kilo Code Inc. and published in the Kilo-Org/kilocode repository under the MIT licence; the licence file names both Kilo Code and “opencode” as copyright holders, since the Kilo CLI is, per its README, a fork of OpenCode extended for the Kilo platform. Since July 2026, Kilo Code has, per the official announcement, been part of Anaconda; per the vendor, the product remains available for individual developers, teams and organizations and is additionally being integrated into the Anaconda platform. Kilo Code is available as a VS Code extension, a JetBrains plugin, a CLI, and a browser-based cloud agent. The CLI is installed via, for example, npm install -g @kilocode/cli or brew install Kilo-Org/tap/kilo, and started by running kilo inside a project directory; the IDE extensions are installed from their respective marketplaces.
Kilo Code as an AI harness
The agent loop evaluates tool requests and writes results back into the context - interactively in the IDE or terminal, or headless via kilo run --auto for CI/CD. Built-in agents are Code (default, full access), Plan (architecture and implementation plans), Ask (read-only), Debug (troubleshooting), and Review (checks changes for performance, security, style and test coverage); custom agents can be defined as markdown files with YAML frontmatter in .kilo/agents/. Tools are extended through local (STDIO) or remote (SSE) MCP servers curated in the Kilo Marketplace, and through skills following the open Agent Skills format (SKILL.md). Permissions are set per action to allow, ask or deny; a checkpoint system automatically saves the file state at the boundaries of every model call and allows files and conversation history to be rolled back to a given message.
Models and model access
Per the vendor, more than 500 models from various providers are available, including Claude, GPT and Gemini - billed at each provider’s own pricing with no markup, with the option to switch models mid-task. Custom OpenAI-compatible endpoints are added via Settings -> Providers -> “Custom provider” with a base URL, API key and model ID, or directly in kilo.jsonc. That means Kilo Code can be run through the AI Gateway, provided the gateway is registered as a custom OpenAI-compatible provider; for Azure OpenAI deployments, the documentation explicitly recommends the native azure provider instead of the custom-provider route.
Enterprise use: data protection and governance
Per the vendor, Kilo Enterprise offers single sign-on through common identity providers, role-based access control (including Owner, Admin, Billing Manager, Member), and audit logs viewable by organization Owners in the Enterprise Dashboard. Providers and models are enabled by default; administrators can selectively block individual providers or models for the whole organization. For paid plans, the vendor states no data retention for prompts and outputs, plus an optional EU data residency for selected processing paths, for example via EU providers such as Inceptron in the Kilo Gateway or through BYOK; a data processing agreement is available for Enterprise customers. Regardless of the harness used, source code and prompts go to whichever model provider is chosen, and that flow belongs in the record of processing activities beforehand.
Orchestrating Kilo Code with herdr
Per herdr’s documentation, the Kilo Code CLI follows the same integration rule as OpenCode: herdr uses a lifecycle plugin when one is installed, falling back to the screen manifest otherwise, and reports both state and session. Details and the full list of supported harnesses are on the herdr page.
Where it fits: developer harness and enterprise harness
Kilo Code is a harness for developers working in an IDE or terminal - for business users, CompanyGPT is the matching harness, with MCP tools, enterprise knowledge, and roles sourced from Entra ID or Keycloak. The AI Gateway provides the shared cost and governance layer across both. Our overview of all AI harnesses covers further coding agents and orchestrators. We help organizations integrate Kilo Code and comparable harnesses cleanly into existing identity, cost and compliance structures.