Skip to content
MetaioLab MetaioLab Docs
Open metaiolab.com

Cherry Studio

Best for: users who want a polished desktop AI app with a simple settings UI.

Why it is a strong fit

Section titled "Why it is a strong fit"

Cherry Studio is one of the easiest AI clients to recommend because:

  • it has a clean desktop experience
  • it supports custom AI providers
  • it has an official NewAPI provider guide
  • it also supports code-agent workflows for tools like Claude Code and Codex
Section titled "Recommended MetaioLab setup"

Use MetaioLab as an OpenAI-compatible provider unless you have a very specific reason to do otherwise.

What to enter

Section titled "What to enter"
  • Provider type: OpenAI
  • API key: your MetaioLab key
  • Base URL: your MetaioLab API base URL
  • Models: fetch automatically if available, or add manually

The easiest path

Section titled "The easiest path"

Cherry Studio has an official NewAPI guide. That is the closest match to a MetaioLab-style deployment built on new-api.

Simple steps

Section titled "Simple steps"
  1. Open Settings
  2. Go to Model Services
  3. Click Add
  4. Choose OpenAI as the provider type
  5. Paste your MetaioLab API key
  6. Paste your MetaioLab base URL
  7. Add or fetch models
  8. Turn the provider on

Very important detail

Section titled "Very important detail"

Cherry Studio’s docs also note a special case for some custom OpenAI providers:

  • if the client’s automatic endpoint appending causes trouble, a trailing # can stop Cherry Studio from auto-appending /chat/completions

Most users should start without that trick. Only use it if your specific endpoint behavior requires it.

What to tell customers

Section titled "What to tell customers"

Use wording like this:

Open Cherry Studio, add an OpenAI-compatible provider, paste your MetaioLab API key, then set the MetaioLab API base URL and choose a model.

Common mistakes

Section titled "Common mistakes"

Wrong base URL

Section titled "Wrong base URL"

Do not paste a docs URL. Use the actual API base URL.

No model selected

Section titled "No model selected"

Even when the key is correct, the client still needs a valid model entry.

Section titled "Stream-related errors"

If your endpoint is older or non-standard, check whether stream behavior needs to be adjusted.

References

Section titled "References"