Cherry Studio
Cherry Studio is a desktop AI client (macOS,
Windows, Linux) with a built-in custom-provider option. It runs natively, so it
reaches the proxy on localhost directly.
Before you start
Do the shared setup from How-to guides first: the proxy must be running and funded.
Connect it
Add a provider
Open Settings (the gear icon), go to Model Providers, and click
+ Add below the provider list. Give it a name like ZeroSignal and choose
OpenAI as the provider type.
Fill in the connection
- API Key: anything non-empty — the proxy ignores it. (Cherry Studio's Check button may probe the key; if it complains, ignore it and add a model in the next step — the actual chat request is what matters.)
- API Address / Base URL:
http://localhost:9376/v1
Add a model
Cherry Studio doesn't auto-load the catalog for custom providers — click
+ Add under Models and enter a model id from the network (e.g.
qwen3-coder; list them with curl http://localhost:9376/v1/models). Add as
many as you want to switch between.
Enable it and chat
Flip the enable switch on the ZeroSignal provider in the list, then start a new chat, pick a model, and send a message. The first reply confirms the path works end to end.
Troubleshooting
| Symptom | Fix |
|---|---|
| Check button says the key is invalid | Expected — the proxy doesn't use keys. Add a model and try a real chat; that's the true test. |
| 404 / "not found" on send | The base URL is off. Make sure the API Address is exactly http://localhost:9376/v1. |
| Nothing happens on send | Confirm the proxy is up (curl http://localhost:9376/healthz) and the provider's enable switch is on. |
wallet_unfunded / payment errors | zs-proxy fund — see Wallet & funding. |
What's next
- Pricing — what a message costs.
- Configuration — spend caps and the output ceiling.
- How-to guides — connect another app.