How to add a custom API host in Cherry Studio
Cherry Studio lets you add a provider that speaks the OpenAI protocol. Create the provider, set the API host to https://api.quickrouter.homes/v1 and paste your gateway key, then add the model ids you want in the chat model picker.
Last updated: October 2026What you need before you start
Cherry Studio is a desktop client, so the key lives in the app rather than in a dotfile - which makes cleaning up after a shared machine more important, not less.
- A QuickRouter account with an active plan and an API key from the console.
- The API host https://api.quickrouter.homes/v1 - Cherry Studio expects the OpenAI-style path.
- The model ids you plan to use, spelled as the gateway serves them.
- A Cherry Studio version recent enough to support custom OpenAI-compatible providers.
Add the provider
Everything happens in one settings screen. Set the host and key first, then add models, because the model list is what the chat picker reads.
- 01
Open the provider list
In Cherry Studio, open Settings and go to Model providers. Add a new provider and choose the OpenAI-compatible type.
- 02
Fill in the host and key
Paste the full path, /v1 included, and your gateway key.
API host: https://api.quickrouter.homes/v1 API key: sk-qr-your-key - 03
Add the models
Add each model id you want to call. A model that is reachable through the gateway is invisible in the chat picker until it is listed here.
gpt-5 claude-sonnet-4-5 gemini-2-5-pro - 04
Send a test message
Pick the new provider in a chat, send one message and confirm the request appears in the gateway console usage view.
Errors you are most likely to hit
Desktop clients hide the underlying HTTP status behind a generic toast, so test the same key with a curl request before you start changing settings at random.
| Symptom | Cause | Fix |
|---|---|---|
| The provider shows no models | No model ids were added to the provider | Add the ids manually - the gateway does not push a model list into the picker |
| 404 from the API | The host field holds https://api.quickrouter.homes without /v1 | Set the API host to https://api.quickrouter.homes/v1 |
| 401 or invalid key | The key was trimmed when copied from the console | Paste again and avoid editing the middle of the string |
| Streaming stops mid-answer | A local proxy or VPN is buffering the connection | Disable the interceptor and retry the same prompt |
| Images fail on a text model | The selected model does not accept image input | Switch to a vision-capable model id from the catalogue |
Which models should Cherry Studio use?
A desktop client is usually a comparison surface, so keep two or three models from different vendors in the list rather than one per vendor.
Frequently asked questions
Does Cherry Studio support OpenAI-compatible providers?+
Yes. Add a provider and choose the OpenAI-compatible type, which is what the gateway exposes.
Why are no models showing after I add the key?+
Cherry Studio reads the provider model list from your configuration, not from the endpoint, so you add the model ids yourself.
Should the API host include /v1?+
Yes. Cherry Studio expects the OpenAI-style path: https://api.quickrouter.homes/v1.
Is my key stored locally?+
Cherry Studio keeps provider credentials on the machine where it runs. Treat a shared or borrowed machine as a machine you need to clear before you hand it back.
One key covers every model in this guide
Create an account, pick a plan and copy an API key. Gateway plans start at $19 per month and model usage is billed at upstream list prices.