Available providers

This page lists every provider type, with its fields and its body in a file. servflowai resource provider describe --all prints the same catalog, with the models each type suggests. For what a provider is, see Providers. To add one in the dashboard, see Manage LLM providers. In a body, config holds the type’s fields as plain values. api_key is the key itself, not the name of a stored secret. The commands that take a body are in the CLI reference.

Type	Name in the dashboard	Needs an endpoint
`claude`	Claude	No
`openai`	OpenAI	No
`openrouter`	OpenRouter	No. It has a default.
`litellm`	LiteLLM	Yes
`qwen`	Qwen	No. It has a default.

Types

claude

Anthropic Claude models.

Key	Label	Type	Required	Description
`api_key`	API Key	password	Yes	The Anthropic API key.
`model`	Model	string	No	The model to call. Defaults to the type’s suggested model.

{
  "name": "Claude production",
  "type": "claude",
  "config": { "api_key": "sk-ant-...", "model": "claude-sonnet-5" }
}

openai

OpenAI models.

Key	Label	Type	Required	Description
`api_key`	API Key	password	Yes	The OpenAI API key.
`model`	Model	string	No	The model to call. Defaults to the type’s suggested model.

{
  "name": "OpenAI production",
  "type": "openai",
  "config": { "api_key": "sk-...", "model": "gpt-4.1" }
}

openrouter

OpenRouter, one key for many companies’ models through an OpenAI-compatible endpoint.

Key	Label	Type	Required	Default	Description
`api_key`	API Key	password	Yes	none	The OpenRouter API key.
`defaultEndpoint`	Endpoint	string	No	`https://openrouter.ai/api/v1/chat/completions`	The chat completions URL.
`model`	Model	string	No	The type’s suggested model	The model to call, as `company/model`.
`reasoning`	Reasoning	select	No	`default`	One of `default`, `none`, `minimal`, `low`, `medium`, `high`. `default` sends no reasoning setting and lets the model decide. Any other value is sent as the reasoning effort.

{
  "name": "OpenRouter",
  "type": "openrouter",
  "config": {
    "api_key": "<your OpenRouter key>",
    "model": "anthropic/claude-sonnet-4.5",
    "reasoning": "low"
  }
}

litellm

A LiteLLM proxy, a self-hosted OpenAI-compatible gateway.

Key	Label	Type	Required	Description
`api_key`	API Key	password	Yes	The LiteLLM API key.
`defaultEndpoint`	Base URL	string	Yes	The proxy’s chat completions URL, such as `http://localhost:4000/v1/chat/completions`.
`model`	Model	string	No	The model to call. Defaults to the type’s suggested model.

{
  "name": "LiteLLM proxy",
  "type": "litellm",
  "config": {
    "api_key": "<your LiteLLM key>",
    "defaultEndpoint": "http://localhost:4000/v1/chat/completions",
    "model": "gpt-4o"
  }
}

qwen

Alibaba Qwen models.

Key	Label	Type	Required	Default	Description
`api_key`	API Key	password	Yes	none	The Qwen API key.
`defaultEndpoint`	Endpoint	string	No	`https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions`	The chat completions URL.
`model`	Model	string	No	The type’s suggested model	The model to call.

{
  "name": "Qwen",
  "type": "qwen",
  "config": { "api_key": "<your Qwen key>", "model": "qwen-plus" }
}

Providers

What a provider is and why sub-agents share one.

Manage LLM providers

Add, change, and delete a provider in the dashboard.

CLI reference

The commands that create and update a provider.

Getting started

Guides

Concepts

References

Types

claude

openai

openrouter

litellm

qwen

Providers

Manage LLM providers

CLI reference

​Types

​claude

​openai

​openrouter

​litellm

​qwen

​Related

Providers

Manage LLM providers

CLI reference

Types

claude

openai

openrouter

litellm

qwen

Related