Chat Playground
Want to chat with models directly inside One API Pro to compare them? That's what this is.
What it is
A built-in chat page at /chat. You can:
- Switch models inside the same conversation to compare answers
- Iterate on prompts without re-running CLI commands
- Demo LLM behavior to teammates / customers
How to open
No extra setup — just visit /chat after logging in. If the page is blank, it means the admin hasn't set ChatLink (it's empty by default).
Usage
Same as ChatGPT / Claude:
- Pick a model (dropdown lists every model you can use)
- Type your prompt
- See the response
Tip: open two tabs with different models for the same prompt to compare side-by-side.
vs. third-party Chat UIs
The Chat Playground is a simple built-in page. For a richer experience (multiple conversations, file uploads, plugins, etc.), use a third-party Chat UI connected via your API Key.
See "Third-party Chat UI" below.
Third-party Chat UI
The recommended path: use Lobe Chat / NextChat / Open WebUI and connect via One API Pro's sk- key.
| Third-party UI | How to configure |
|---|---|
| Lobe Chat | Pick "OpenAI compatible"; Base URL = http://<your-host>:3000; API Key = sk-… |
| NextChat | Same |
| Open WebUI | Same; supports streaming, tools, multimodal |
Steps:
- Personal Center → Tokens → create an API Key (recommended: "never expires + unrestricted models")
- In the third-party UI, choose OpenAI-compatible; Base URL = your One API Pro (
http://localhost:3000or your domain); API Key = thesk-… - Models configured in Model Price will appear in the UI
FAQ
/chatis blank: admin hasn't setChatLink. Ask the admin to configure System Settings, or just use a third-party UI.- Calls return 401: create a "never expires + unrestricted" API Key and try again.
- Third-party UI shows "no models": check its Base URL points to One API Pro and the API Key is valid.
- CORS error: deploy the UI and One API Pro on the same origin (one reverse proxy), or add
Access-Control-Allow-Originin the reverse proxy.