DeepSeek API and MCP: how to connect the code models to your project

DeepSeek is a family of chat models for code and text. NeuralSpace offers the following models: DeepSeek V4 Flash, DeepSeek V4 Pro. They all run through a single public API, and there is a ready MCP server for AI agents. The key is issued on the API keys page.

What DeepSeek can do

DeepSeek models are built for vibe coding and text work. V4 Flash is the fast, affordable version for high-volume tasks, V4 Pro is more powerful for complex code and reasoning. Both understand and write code in many languages, explain it and hold conversations.

API endpoint

Send a request to POST /v1/chat/completions. Example: {"model": "deepseek-v4-pro", "messages": [{"role": "user", "content": "Write a function"}]}. Authorization uses an API key in the x-api-key header. Available model ids: deepseek-v4-flash, deepseek-v4-pro.

MCP tools

The NeuralSpace MCP server is available at POST /mcp (Streamable HTTP). To work with DeepSeek, an agent uses the tools: chat, list_models, get_balance. This lets Claude Code, Cursor and other MCP clients use the site models directly.

How to connect

Sign up on NeuralSpace, open the API keys page and create an nsk-… key. Then pass it in the request header or in your MCP client settings. Full documentation is on the the /v1/docs page.

FAQ

How does Flash differ from Pro?

Flash is faster and cheaper, Pro is more powerful for complex code and reasoning.

Is it good for vibe coding?

Yes, both versions are tuned for code generation and explanation.

What format is the request?

OpenAI-compatible: POST /v1/chat/completions.