Skip to main content
Providers are LLM services that power the nanobot agent. Each provider requires an API key and has specific configuration options.

​
Provider Auto-Detection

Nanobot automatically detects the correct provider based on your model name: 
{
  "agents": {
    "defaults": {
      "model": "anthropic/claude-opus-4-5",
      "provider": "auto"
    }
  }
}
The provider field can be:
  • "auto" — Auto-detect from model name (recommended)
  • Provider name — Force a specific provider (e.g. "openrouter", "anthropic")

​
Provider Configuration

Each provider has three common fields:
​
providers.{name}.apiKey
string
required
API key for authentication. Required for all providers except OAuth-based ones.
​
providers.{name}.apiBase
string
Custom API base URL. Overrides the default endpoint.
​
providers.{name}.extraHeaders
object
Custom HTTP headers to include in API requests (e.g. for gateway authentication).

​
Available Providers

​
Gateway Providers

Gateway providers can route any model and provide access to multiple LLM backends.
Global LLM gateway with access to all major models.Configuration:
{
  "providers": {
    "openrouter": {
      "apiKey": "sk-or-v1-xxx"
    }
  }
}
Get API Key: openrouter.ai/keysDefault API Base: https://openrouter.ai/api/v1Model Prefix: openrouter/ (auto-added)Features:
  • Access to 200+ models
  • Prompt caching support
  • Auto-detected by sk-or- key prefix
OpenAI-compatible API gateway.Configuration:
{
  "providers": {
    "aihubmix": {
      "apiKey": "your-key",
      "extraHeaders": {
        "APP-Code": "your-app-code"
      }
    }
  }
}
Get API Key: aihubmix.comDefault API Base: https://aihubmix.com/v1Note: Strips existing model prefixes before routing (e.g. anthropic/claude-3 → claude-3 → openai/claude-3).
Chinese LLM gateway with OpenAI-compatible API.Configuration:
{
  "providers": {
    "siliconflow": {
      "apiKey": "your-key"
    }
  }
}
Get API Key: siliconflow.cnDefault API Base: https://api.siliconflow.cn/v1
ByteDance’s LLM gateway.Configuration:
{
  "providers": {
    "volcengine": {
      "apiKey": "your-key"
    }
  }
}
Get API Key: volcengine.comDefault API Base: https://ark.cn-beijing.volces.com/api/v3Coding Plan: For VolcEngine coding plan, set "apiBase": "https://ark.cn-beijing.volces.com/api/coding/v3"

​
Standard Providers

Direct connections to individual LLM providers.
Direct access to Claude models.Configuration:
{
  "providers": {
    "anthropic": {
      "apiKey": "sk-ant-xxx"
    }
  },
  "agents": {
    "defaults": {
      "model": "claude-opus-4-5"
    }
  }
}
Get API Key: console.anthropic.comKeywords: anthropic, claudeFeatures:
  • Native Claude models
  • Prompt caching support
  • No prefix required
Direct access to GPT models.Configuration:
{
  "providers": {
    "openai": {
      "apiKey": "sk-xxx"
    }
  },
  "agents": {
    "defaults": {
      "model": "gpt-4o"
    }
  }
}
Get API Key: platform.openai.comKeywords: openai, gpt
DeepSeek models (China-based).Configuration:
{
  "providers": {
    "deepseek": {
      "apiKey": "sk-xxx"
    }
  },
  "agents": {
    "defaults": {
      "model": "deepseek-chat"
    }
  }
}
Get API Key: platform.deepseek.comModel Prefix: deepseek/ (auto-added)Keywords: deepseek
Google’s Gemini models.Configuration:
{
  "providers": {
    "gemini": {
      "apiKey": "your-key"
    }
  },
  "agents": {
    "defaults": {
      "model": "gemini-pro"
    }
  }
}
Get API Key: aistudio.google.comModel Prefix: gemini/ (auto-added)
Moonshot AI’s Kimi models.Configuration:
{
  "providers": {
    "moonshot": {
      "apiKey": "sk-xxx",
      "apiBase": "https://api.moonshot.ai/v1"
    }
  },
  "agents": {
    "defaults": {
      "model": "kimi-k2.5"
    }
  }
}
Get API Key: platform.moonshot.cnModel Prefix: moonshot/ (auto-added)Default API Base: https://api.moonshot.ai/v1 (international) or https://api.moonshot.cn/v1 (China)Note: Kimi K2.5 requires temperature >= 1.0 (enforced automatically).
Zhipu AI’s GLM models.Configuration:
{
  "providers": {
    "zhipu": {
      "apiKey": "your-key"
    }
  },
  "agents": {
    "defaults": {
      "model": "glm-4"
    }
  }
}
Get API Key: open.bigmodel.cnModel Prefix: zai/ (auto-added)Keywords: zhipu, glm, zaiCoding Plan: For Zhipu coding plan, set "apiBase": "https://open.bigmodel.cn/api/coding/paas/v4"
Alibaba Cloud’s Qwen models.Configuration:
{
  "providers": {
    "dashscope": {
      "apiKey": "sk-xxx"
    }
  },
  "agents": {
    "defaults": {
      "model": "qwen-max"
    }
  }
}
Get API Key: dashscope.console.aliyun.comModel Prefix: dashscope/ (auto-added)Keywords: qwen, dashscope
MiniMax models.Configuration:
{
  "providers": {
    "minimax": {
      "apiKey": "your-key"
    }
  },
  "agents": {
    "defaults": {
      "model": "MiniMax-M2.1"
    }
  }
}
Get API Key: platform.minimaxi.comDefault API Base: https://api.minimax.io/v1 (international) or https://api.minimaxi.com/v1 (China)Note: For mainland China API, set "apiBase": "https://api.minimaxi.com/v1"
High-speed inference for voice transcription (Whisper) and LLM.Configuration:
{
  "providers": {
    "groq": {
      "apiKey": "gsk_xxx"
    }
  }
}
Get API Key: console.groq.comUse Case: Mainly used for free Whisper voice transcription in Telegram/Discord.

​
OAuth Providers

Providers that use OAuth authentication instead of API keys.
ChatGPT Codex models via OAuth (requires ChatGPT Plus/Pro).Login:
nanobot provider login openai-codex
Configuration:
{
  "agents": {
    "defaults": {
      "model": "openai-codex/gpt-5.1-codex"
    }
  }
}
Requirements: ChatGPT Plus or Pro subscription
GitHub Copilot models via OAuth.Login:
nanobot provider login github-copilot
Configuration:
{
  "agents": {
    "defaults": {
      "model": "github_copilot/gpt-4o"
    }
  }
}

​
Local/Self-Hosted

Self-hosted models via vLLM or any OpenAI-compatible server.Start Server:
vllm serve meta-llama/Llama-3.1-8B-Instruct --port 8000
Configuration:
{
  "providers": {
    "vllm": {
      "apiKey": "dummy",
      "apiBase": "http://localhost:8000/v1"
    }
  },
  "agents": {
    "defaults": {
      "model": "meta-llama/Llama-3.1-8B-Instruct"
    }
  }
}
Note: API key can be any non-empty string for local servers.
Connect to any OpenAI-compatible endpoint (LM Studio, llama.cpp, Together AI, Fireworks, Azure OpenAI).Configuration:
{
  "providers": {
    "custom": {
      "apiKey": "your-api-key",
      "apiBase": "https://api.your-provider.com/v1"
    }
  },
  "agents": {
    "defaults": {
      "model": "your-model-name"
    }
  }
}
Note: Model name is passed as-is without LiteLLM transformation. For local servers without authentication, set "apiKey": "no-key".

​
Example: Multi-Provider Setup

{
  "providers": {
    "openrouter": {
      "apiKey": "sk-or-v1-xxx"
    },
    "anthropic": {
      "apiKey": "sk-ant-xxx"
    },
    "groq": {
      "apiKey": "gsk_xxx"
    }
  },
  "agents": {
    "defaults": {
      "model": "anthropic/claude-opus-4-5",
      "provider": "openrouter"
    }
  }
}
This setup:
  • Uses OpenRouter as the primary LLM gateway
  • Has Anthropic configured for direct Claude access
  • Enables Groq for voice transcription

​
Provider Selection Priority

  1. Explicit provider prefix in model name: github-copilot/model always uses GitHub Copilot
  2. Model name keywords: claude-opus auto-matches Anthropic
  3. Gateway detection: API key prefix (sk-or-) or base URL keyword
  4. Fallback: First configured provider with an API key

​
Next Steps

Channels

Connect nanobot to chat platforms

Security

Configure access controls and sandboxing

Build docs developers (and LLMs) love

Get started for freeTalk to us