Skills matrix
| Skill | Status | Delivery | Maturity | Evidence |
|---|---|---|---|---|
| Code generation | Partial | Model-dependent | Model-limited | Not publicly documented |
| Refactoring | Partial | Model-dependent | Model-limited | Not publicly documented |
| Multi-file reasoning | Partial | Model-dependent | Model-limited | Not publicly documented |
| Test generation | Partial | Model-dependent | Model-limited | Not publicly documented |
| Static analysis support | Partial | Model-dependent | Model-limited | Not publicly documented |
| Codebase indexing | Partial | Model-dependent | Model-limited | Not publicly documented |
| Semantic retrieval | Partial | Model-dependent | Model-limited | Not publicly documented |
| Memory retention across sessions | Partial | Model-dependent | Model-limited | Not publicly documented |
| Context injection control | Partial | Model-dependent | Model-limited | Not publicly documented |
| Inline editing | Partial | Model-dependent | Model-limited | Not publicly documented |
| Chat-first interaction | Partial | Model-dependent | Model-limited | Not publicly documented |
| File diff preview | Partial | Model-dependent | Model-limited | Not publicly documented |
| Git integration | Partial | Model-dependent | Model-limited | Not publicly documented |
| Terminal / command execution | Partial | Model-dependent | Model-limited | Not publicly documented |
| Model switching | Partial | Model-dependent | Model-limited | Not publicly documented |
| Multi-model orchestration | Partial | Model-dependent | Model-limited | Not publicly documented |
| Prompt augmentation layer | Partial | Model-dependent | Model-limited | Not publicly documented |
| Agent loop execution | Partial | Model-dependent | Model-limited | Not publicly documented |
| Local model support | Partial | Model-dependent | Model-limited | Not publicly documented |
| Offline capability | Partial | Model-dependent | Model-limited | Not publicly documented |
| Enterprise policy control | Partial | Model-dependent | Model-limited | Not publicly documented |
Why it matters
- 25 trillion tokens processed validates production-scale reliability — this isn't a demo project; it handles real developer workloads at massive scale.
- #1 on OpenRouter means Kilo's model routing and prompt engineering have been battle-tested across millions of real coding tasks.
- Apache 2.0 open source with 15K+ stars provides the largest open-source agentic coding agent community — extensive community plugins, documentation, and support.
- BYOK model gives full cost and model control — use any OpenRouter-compatible model, switch between fast and capable models per task type.
Key capabilities
- Agentic task execution: Multi-step autonomous task completion — read files, write code, run commands, debug iteratively.
- Code completion: Inline AI completions in VS Code.
- AI chat: In-editor chat with full codebase context.
- Model flexibility: Works with any OpenAI-compatible API via OpenRouter — GPT, Claude, Gemini, Llama.
- VS Code extension: Native VS Code integration with sidebar chat and inline completion.
- Custom instructions: Project-level and global custom instructions for AI behavior.
- MCP support: Model Context Protocol integration for extended tool capabilities.
- Open source: Apache 2.0; fork, extend, self-host.
Technical notes
- License: Apache 2.0
- GitHub: github.com/kilocode-ai/kilocode (check actual repo URL)
- Stars: 15,694
- Platform: VS Code extension + web
- LLM: BYOK — OpenRouter, OpenAI, Anthropic, any compatible API
- Usage: 1.5M+ developers; 25T+ tokens processed
- OpenRouter rank: #1 by usage volume
Ideal for
- Developers who want a fully open-source agentic coding agent with a large community and proven scale.
- Teams requiring full control over model selection and costs through BYOK/OpenRouter integration.
- Organizations needing Apache 2.0 licensing for commercial deployment or internal tooling without restrictions.
Not ideal for
- Teams wanting a fully managed, zero-configuration experience — Cursor or GitHub Copilot are more plug-and-play.
- Non-VS Code editors — Kilo is primarily VS Code; other editors may have limited support.
- Users without LLM API keys — Kilo requires your own API credentials.