Desktop Chat Clients

Replace {baseUrl} with your platform API root, without a trailing / or /v1, and sk-your-api-key with your model API key. Model IDs and interface labels may vary by client version and account.

Cherry Studio

Cherry Studio is a desktop AI client with built-in assistants, parallel conversations with multiple models, knowledge bases, and image generation.

• Website
• Download

Configuration

1. Copy your key from Console → Tokens.
2. Open Cherry Studio → Settings → Model Services → Add.
3. Enter the following settings:

Setting	Value
Provider name	MRouter (or another name)
Provider type	OpenAI or NewAPI
API key	Your sk- key
API URL	{baseUrl}/v1

4. Add model IDs in the provider's Models section, for example:
   • gpt-5.2 for general chat.
   • claude-sonnet-4-5-20250929 for code and long-form text.
   • gemini-2.5-pro for multimodal tasks.
   • deepseek-v4-pro for cost-sensitive use.
5. Use the provider's Check button to retrieve available models through /v1/models, where supported by your version.
6. Return to chat and select the model from the dropdown.

Image generation: add a model that supports images, such as gemini-3-pro-image or Doubao-Seedream-5.0-pro, select it, and describe the image you want.

Troubleshooting: if the connection check fails, check the URL first. The API URL must include /v1 exactly once.

AQBot

AQBot is a local-first, open-source desktop AI workspace licensed under AGPL-3.0. Data and files remain on your machine.

• Website
• Downloads

Manual configuration

1. Open Settings → Providers → Add Provider.
2. Enter:

Setting	Value
Provider name	MRouter
Provider type	OpenAI
API URL	{baseUrl} — omit /v1 because AQBot adds it.
API key	sk-your-api-key

3. Select Sync Models to retrieve available models.
4. Return to chat and select a model.

For the OpenAI provider, AQBot automatically appends /v1. Versions described in the source guide avoid adding it twice when it is already present, and allow a trailing ! to disable automatic completion. Using the root URL is the recommended configuration.

Optional native protocols

You can add more providers with the same key:

Use case	Provider type	API URL
Chat Completions	OpenAI	{baseUrl}
Native Claude /v1/messages	Anthropic	{baseUrl}
Native Gemini	Gemini	{baseUrl}

LobeChat

Docker deployment

docker run -d -p 3210:3210 \
  -e OPENAI_API_KEY=sk-your-api-key \
  -e OPENAI_PROXY_URL={baseUrl}/v1 \
  -e ACCESS_CODE=your-access-password \
  lobehub/lobe-chat

Vercel: set the OPENAI_API_KEY and OPENAI_PROXY_URL={baseUrl}/v1 environment variables.

In the UI: open Settings → Language Models → OpenAI, enter the key and proxy URL ({baseUrl}/v1), then add model IDs to the model list.

NextChat (ChatGPT-Next-Web)

1. Open NextChat settings.
2. Enable Custom API.
3. Enter:

Setting	Value
API URL	{baseUrl} — omit /v1.
API key	sk-your-api-key
Custom model names	gpt-5.2,claude-sonnet-4-5-20250929,deepseek-v4-pro (comma-separated)

NextChat appends /v1/chat/completions, so enter only the platform root URL.

Open WebUI

Docker deployment

docker run -d -p 3000:8080 \
  -e OPENAI_API_BASE_URL={baseUrl}/v1 \
  -e OPENAI_API_KEY=sk-your-api-key \
  -v open-webui:/app/backend/data \
  --name open-webui \
  ghcr.io/open-webui/open-webui:main

In the UI: open Admin Settings → Connections → OpenAI, enter the URL and key, refresh the model list, then save.