# AI Providers

> Configure and connect different AI model providers to Kilo Code

- 网址：https://funcoding.ai/agents/kilo-code/ai-providers/
- 来源：Kilo Code 官方文档原文（英文），MIT 许可，同步于 2026-10-11
- 官方原文：https://kilo.ai/docs/ai-providers

---
Kilo Code supports a wide variety of AI providers, giving you flexibility in how you power your AI-assisted development workflow. Choose from cloud providers, local models, or AI gateways based on your needs.

## Getting Started

The fastest way to get started is with **Kilo Code's built-in provider**, which requires no configuration. Simply sign in and start coding.

For users who want to use their own API keys or need specific models, we support over 30 providers.

## Provider Categories

### Cloud Providers

Major AI companies offering powerful models via API:

- **[Anthropic](https://funcoding.ai/agents/kilo-code/ai-providers/anthropic/)** - Claude models (Claude 4, Claude 3.5 Sonnet, etc.)
- **[OpenAI](https://funcoding.ai/agents/kilo-code/ai-providers/openai/)** - GPT-4, GPT-4o, o1, and more
- **[Google Gemini](https://funcoding.ai/agents/kilo-code/ai-providers/gemini/)** - Gemini Pro, Gemini Ultra
- **[Google Vertex AI](https://funcoding.ai/agents/kilo-code/ai-providers/vertex/)** - Google Cloud-hosted Gemini and partner models
- **[AWS Bedrock](https://funcoding.ai/agents/kilo-code/ai-providers/bedrock/)** - AWS-hosted foundation models
- **[Alibaba Cloud](https://funcoding.ai/agents/kilo-code/ai-providers/alibaba/)** - DashScope and Qwen models through Model Studio
- **[Cloudflare](https://funcoding.ai/agents/kilo-code/ai-providers/cloudflare/)** - Workers AI and Cloudflare AI Gateway
- **[DeepSeek](https://funcoding.ai/agents/kilo-code/ai-providers/deepseek/)** - DeepSeek V3., R1
- **[Mistral](https://funcoding.ai/agents/kilo-code/ai-providers/mistral/)** - Mistral Large, Codestral
- **[Nebius Token Factory](https://funcoding.ai/agents/kilo-code/ai-providers/nebius/)** - Model inference on Nebius infrastructure
- **[Poolside](https://funcoding.ai/agents/kilo-code/ai-providers/poolside/)** - Laguna models

### Local & Self-Hosted

Run models on your own hardware for privacy and offline use:

- **[Atomic Chat](https://funcoding.ai/agents/kilo-code/ai-providers/atomic-chat/)** - Local models with TurboQuant inference and auto-discovery in Kilo Code
- **[Anaconda Desktop](https://funcoding.ai/agents/kilo-code/ai-providers/anaconda-desktop/)** - Discover and connect to a local text-generation model server
- **[Ollama](https://funcoding.ai/agents/kilo-code/ai-providers/ollama/)** - Easy local model management
- **[LM Studio](https://funcoding.ai/agents/kilo-code/ai-providers/lmstudio/)** - Desktop app for local models
- **[OpenAI Compatible](https://funcoding.ai/agents/kilo-code/ai-providers/openai-compatible/)** - Any OpenAI-compatible endpoint

### AI Gateways

Route requests through unified APIs with additional features:

- **[OpenRouter](https://funcoding.ai/agents/kilo-code/ai-providers/openrouter/)** - Access multiple providers through one API
- **[TrustedRouter](https://funcoding.ai/agents/kilo-code/ai-providers/trustedrouter/)** - OpenAI-compatible access to attested routing, ZDR routing, and E2E encrypted model routes
- **[Requesty](https://funcoding.ai/agents/kilo-code/ai-providers/requesty/)** - Smart routing and fallbacks
- **[DaoXE](https://funcoding.ai/agents/kilo-code/ai-providers/daoxe/)** - Connect multiple model families through one API
- **[Cloudflare AI Gateway](https://funcoding.ai/agents/kilo-code/ai-providers/cloudflare/)** - Route providers through your Cloudflare account
- **[Eden AI](https://funcoding.ai/agents/kilo-code/ai-providers/edenai/)** - EU-based gateway with one key across vendors

## Choosing a Provider

| Priority | Recommended Provider |
|---|---|
| Ease of use | [Kilo Code (built-in)](https://funcoding.ai/agents/kilo-code/ai-providers/kilocode/) |
| Best value | Zhipu AI or Mistral |
| Privacy/Offline | Ollama or LM Studio |
| Enterprise | AWS Bedrock or Google Vertex |

## Why Use Multiple Providers?

- **Cost** - Compare pricing across providers for different tasks
- **Reliability** - Backup options when a provider has outages
- **Models** - Access exclusive or specialized models
- **Regional** - Better latency in certain locations

## Disabling Built-in Providers

You can prevent specific providers from loading using `disabled_providers` in your `kilo.json` (or `kilo.jsonc`). This is useful to hide models from built-in or detected providers that you don't intend to use.

```json
{
  "$schema": "https://app.kilo.ai/config.json",
  "disabled_providers": ["kilo", "openai"]
}
```

To allow only specific providers and disable everything else, use `enabled_providers` instead:

```json
{
  "$schema": "https://app.kilo.ai/config.json",
  "enabled_providers": ["anthropic"]
}
```

Both fields accept provider IDs — the lowercase identifier used in the `provider/model` format (e.g. `kilo`, `anthropic`, `openai`, `google`, `groq`).

## Next Steps

- **New to Kilo Code?** Start with the [Kilo Code provider](https://funcoding.ai/agents/kilo-code/ai-providers/kilocode/) - no setup required
- **Have an API key?** Jump to your provider's page for configuration instructions
- **Want to compare?** Check out [Model Selection](https://funcoding.ai/agents/kilo-code/code-with-ai/agents/model-selection/) for guidance on choosing models