import { Callout } from 'nextra/components';

# Model Service Providers

When deploying LobeChat, a rich set of environment variables related to model service providers is provided, allowing you to easily define the model service providers to be enabled in LobeChat.

## OpenAI

### `OPENAI_API_KEY`

- Type: Required
- Description: This is the API key you applied for on the OpenAI account page, you can check it out [here](openai-api-page)
- Default: -
- Example: `sk-xxxxxx...xxxxxx`

### `OPENAI_PROXY_URL`

- Type: Optional
- Description: If you manually configure the OpenAI interface proxy, you can use this configuration item to override the default OpenAI API request base URL
- Default: `https://api.openai.com/v1`
- Example: `https://api.chatanywhere.cn` or `https://aihubmix.com/v1`

<Callout type={'warning'}>
  Please check the request suffix of your proxy service provider. Some proxy service providers may
  add `/v1` to the request suffix, while others may not. If you find that the AI returns an empty
  message during testing, try adding the `/v1` suffix and retry.{' '}
</Callout>

<Callout type={'info'}>
  Whether to fill in `/v1` is closely related to the model service provider. For example, the
  default address of openai is `api.openai.com/v1`. If your proxy forwards the `/v1` interface, you
  can simply fill in `proxy.com`. However, if the model service provider directly forwards the
  `api.openai.com` domain, then you need to add `/v1` to the URL yourself.{' '}
</Callout>

Related discussions:

- [Why is the return value blank after installing Docker, configuring environment variables?](https://github.com/lobehub/lobe-chat/discussions/623)
- [Reasons for errors when using third-party interfaces](https://github.com/lobehub/lobe-chat/discussions/734)
- [No response in chat after filling in the proxy server address](https://github.com/lobehub/lobe-chat/discussions/1065)

### `CUSTOM_MODELS`

- Type: Optional
- Description: Used to control the model list, use `+` to add a model, use `-` to hide a model, use `model_name=display_name` to customize the display name of a model, separated by commas.
- Default: `-`
- Example: `+qwen-7b-chat,+glm-6b,-gpt-3.5-turbo,gpt-4-0125-preview=gpt-4-turbo`

The above example adds `qwen-7b-chat` and `glm-6b` to the model list, removes `gpt-3.5-turbo` from the list, and displays the name of `gpt-4-0125-preview` as `gpt-4-turbo`. If you want to disable all models first and then enable specific models, you can use `-all,+gpt-3.5-turbo`, which means only `gpt-3.5-turbo` will be enabled.

You can find all current model names in [modelProviders](https://github.com/lobehub/lobe-chat/tree/main/src/config/modelProviders).

## Azure OpenAI

If you need to use Azure OpenAI to provide model services, you can refer to the [Deploying with Azure OpenAI](../Deployment/Deploy-with-Azure-OpenAI.en-US.md) section for detailed steps. Here, we will list the environment variables related to Azure OpenAI.

### `API_KEY_SELECT_MODE`

- Type：Optional
- Description：Controls the mode for selecting the API Key when multiple API Keys are available. Currently supports `random` and `turn`.
- Default：`random`
- Example：`random` or `turn`

When using the `random` mode, a random API Key will be selected from the available multiple API Keys.

When using the `turn` mode, the API Keys will be retrieved in a round-robin manner according to the specified order.

### `USE_AZURE_OPENAI`

- Type: Optional
- Description: Set this value to `1` to enable Azure OpenAI configuration
- Default: -
- Example: `1`

### `AZURE_API_KEY`

- Type: Optional
- Description: This is the API key you applied for on the Azure OpenAI account page
- Default: -
- Example: `c55168be3874490ef0565d9779ecd5a6`

### `AZURE_API_VERSION`

- Type: Optional
- Description: The API version of Azure, following the format YYYY-MM-DD
- Default: `2023-08-01-preview`
- Example: `2023-05-15`, refer to [latest version][azure-api-verion-url]

## ZHIPU AI

### `ZHIPU_API_KEY`

- Type: Required
- Description: This is the API key you applied for in the ZHIPU AI service
- Default: -
- Example: `4582d332441a313f5c2ed9824d1798ca.rC8EcTAhgbOuAuVT`

## Moonshot AI

### `MOONSHOT_API_KEY`

- Type: Required
- Description: This is the API key you applied for in the Moonshot AI service
- Default: -
- Example: `Y2xpdGhpMzNhZXNoYjVtdnZjMWc6bXNrLWIxQlk3aDNPaXpBWnc0V1RaMDhSRmRFVlpZUWY=`

## Google AI

### `GOOGLE_API_KEY`

- Type: Required
- Description: This is the API key you applied for in the Google AI Platform to access Google AI services
- Default: -
- Example: `AIraDyDwcw254kwJaGjI9wwaHcdDCS__Vt3xQE`

## AWS Bedrock

### `AWS_ACCESS_KEY_ID`

- Type: Required
- Description: Access key ID for AWS service authentication
- Default: -
- Example: `AKIA5STVRLFSB4S9HWBR`

### `AWS_SECRET_ACCESS_KEY`

- Type: Required
- Description: Key for AWS service authentication
- Default: -
- Example: `Th3vXxLYpuKcv2BARktPSTPxx+jbSiFT6/0w7oEC`

### `AWS_REGION`

- Type: Optional
- Description: Region setting for AWS services
- Default: `us-east-1`
- Example: `us-east-1`

## Ollama

### `OLLAMA_PROXY_URL`

- Type: Optional
- Description: Used to enable the Ollama service, setting this will display optional open-source language models in the language model list and can also specify custom language models
- Default: -
- Example: `http://127.0.0.1:11434/v1`

## Perplexity AI

### `PERPLEXITY_API_KEY`

- Type: Required
- Description: This is the API key you applied from Perplexity AI
- Default: -
- Example: `pplx-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx`

## Anthropic AI

### `ANTHROPIC_API_KEY`

- Type: Required
- Description: This is the API key you applied from Anthropic AI
- Default: -
- Example: `sk-ant-apixx-xxxxxxxxx-xxxxxxxxxxxxxxxxx-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx-xxxxxxxx`

## Mistral AI

### `MISTRAL_API_KEY`

- Type: Required
- Description: This is the API key you applied for in the Mistral AI service
- Default: -
- Example: `xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx=`