DeepSeek V4 Playground
Experience the latest DeepSeek V4 model - generate small apps with one prompt. Powered by EdgeOne Makers Models built-in models.
Highlights
- DeepSeek V4: The latest and most powerful DeepSeek model with enhanced code generation capabilities
- Edgeone Makers Models: EdgeOne Makers Models provides built-in models, and can also host self-paid provider keys to proxy requests, delivering low-latency responses
- Live Preview: Instantly preview code through an interactive sandbox
Environment Variables
| Variable | Required | Description |
|---|---|---|
AI_GATEWAY_BASE_URL | Yes | Gateway base URL. When using Edgeone Makers Models, set it to https://ai-gateway.edgeone.link/v1. |
AI_GATEWAY_API_KEY | Yes | Model gateway API key. Use a Edgeone Makers Models API Key, or the key of any OpenAI-compatible provider. |
About Changing Model Invocations
Tech Stack
- Sandpack for the code sandbox
- Next.js App Router with Tailwind
Supported Models
| Model | Description |
|---|---|
@makers/deepseek-v4-flash | Better suited for scenarios that prioritize low latency, high concurrency, and cost efficiency. |
@makers/deepseek-v4-pro | Better suited for scenarios that prioritize reasoning capability, higher answer quality, and complex task performance. |
Get an OpenAI API-Compatible Service
After deployment, this project provides an OpenAI API-compatible endpoint that you can configure for use in various AI tools and applications:
- Endpoint URL:
https://<your_domain>/v1/chat/completions - Compatible Tools: Any AI tools and applications that support the OpenAI API format
Simply configure the deployed endpoint URL in your AI tools to start using it.