DeepSeek MCP Server
- 349 repo stars
- Updated April 24, 2026
- DMontgomery40/deepseek-mcp-server
DeepSeek MCP Server is a MCP server that exposes official DeepSeek chat, completion, model listing, and balance APIs to MCP clients.
About
DeepSeek MCP Server is the official Model Context Protocol bridge to DeepSeek’s API, aimed at developers who want their coding agent to call DeepSeek without maintaining bespoke HTTP clients. Through stdio npm installation, it exposes chat, text completion, model listing, and account balance capabilities so you can prototype agents, compare models, and sanity-check usage from the same session where you edit code. You only need a DeepSeek API key in DEEPSEEK_API_KEY; transport is local stdio, which matches how Claude Code and Cursor typically register MCP servers. It does not replace prompt design or app architecture—it standardizes provider access. Use it when DeepSeek is your inference backend during Build, or when you want quick model and balance visibility while iterating on agent workflows.
- Official DeepSeek MCP for chat and completion endpoints
- Model listing tools to pick the right DeepSeek model in-agent
- Balance endpoint exposure for spend and quota awareness
- stdio npm transport via deepseek-mcp-server 0.3.1
- Single required secret: DEEPSEEK_API_KEY for upstream requests
DeepSeek MCP Server by the numbers
- Data as of Jul 21, 2026 (Skillselion catalog sync)
claude mcp add --env DEEPSEEK_API_KEY=YOUR_DEEPSEEK_API_KEY deepseek-mcp-server -- npx -y deepseek-mcp-serverAdd your badge
Show developers this MCP server is listed on Skillselion. Paste this into your README.
| repo stars | ★ 349 |
|---|---|
| Package | deepseek-mcp-server |
| Transport | STDIO |
| Auth | Required |
| Last updated | April 24, 2026 |
| Repository | DMontgomery40/deepseek-mcp-server ↗ |
What it does
Wire DeepSeek chat, completions, model discovery, and balance checks into your agent stack through a single official MCP server.
Who is it for?
Best when you're standardizing DeepSeek inside Claude Code, Cursor, or Codex during agent and backend experiments.
Skip if: Skip if you only use Anthropic or OpenAI exclusively with no DeepSeek account.
What you get
Install one MCP server so your agent can chat, complete, list models, and check balance on DeepSeek with a single API key.
- MCP tools for DeepSeek chat and completion calls
- In-agent DeepSeek model catalog queries
- Balance visibility for API usage monitoring during development
By the numbers
- Server version 0.3.1
- npm identifier deepseek-mcp-server
- Capabilities: chat, completion, model listing, balance endpoints
README.md
DeepSeek MCP Server
Model Context Protocol server for the current DeepSeek V4 API.
As of April 24, 2026, DeepSeek's public API reference documents:
POST /chat/completionswithdeepseek-v4-flashanddeepseek-v4-proPOST /beta/completionsfor V4 Pro FIM completionGET /modelsGET /user/balance
This server exposes only those documented API surfaces. It does not ship a V4 monitor, speculative image/video/upload tools, or automatic model substitution.
Tools
chat_completion: DeepSeek V4 chat. Defaults todeepseek-v4-flash. Supportsthinking: { "type": "enabled" | "disabled" },reasoning_effort: "high" | "max", JSON output, function tools, logprobs, streaming, and conversation memory.completion: DeepSeek V4 Pro FIM completion. Defaults todeepseek-v4-pro.list_models: Reads the live DeepSeek model list.get_user_balance: Reads account balance and availability.reset_conversation: Clears an in-memory conversation.list_conversations: Lists in-memory conversation IDs.
Hosted Remote
- URL:
https://deepseek-mcp.ragweld.com/mcp - Auth:
Authorization: Bearer <token>
Codex CLI:
export DEEPSEEK_MCP_AUTH_TOKEN="REPLACE_WITH_TOKEN"
codex mcp add deepseek --url https://deepseek-mcp.ragweld.com/mcp --bearer-token-env-var DEEPSEEK_MCP_AUTH_TOKEN
Claude Code:
export DEEPSEEK_MCP_AUTH_TOKEN="REPLACE_WITH_TOKEN"
claude mcp add --transport http deepseek https://deepseek-mcp.ragweld.com/mcp --header "Authorization: Bearer $DEEPSEEK_MCP_AUTH_TOKEN"
Cursor:
node -e 'const fs=require("fs"),p=process.env.HOME+"/.cursor/mcp.json";let j={mcpServers:{}};try{j=JSON.parse(fs.readFileSync(p,"utf8"))}catch{};j.mcpServers={...(j.mcpServers||{}),deepseek:{url:"https://deepseek-mcp.ragweld.com/mcp",headers:{Authorization:"Bearer ${env:DEEPSEEK_MCP_AUTH_TOKEN}"}}};fs.mkdirSync(process.env.HOME+"/.cursor",{recursive:true});fs.writeFileSync(p,JSON.stringify(j,null,2));'
Local Stdio
DEEPSEEK_API_KEY="REPLACE_WITH_DEEPSEEK_KEY" npx -y deepseek-mcp-server
Docker:
docker pull docker.io/dmontgomery40/deepseek-mcp-server:0.5.0
docker run --rm -i -e DEEPSEEK_API_KEY="REPLACE_WITH_DEEPSEEK_KEY" docker.io/dmontgomery40/deepseek-mcp-server:0.5.0
Environment
Required:
DEEPSEEK_API_KEY=your-api-key
Optional:
DEEPSEEK_BASE_URL=https://api.deepseek.com
DEEPSEEK_REQUEST_TIMEOUT_MS=120000
DEEPSEEK_DEFAULT_MODEL=deepseek-v4-flash
MCP_TRANSPORT=stdio
MCP_HTTP_HOST=127.0.0.1
MCP_HTTP_PORT=3001
MCP_HTTP_PATH=/mcp
MCP_HTTP_STATEFUL_SESSION=false
CONVERSATION_MAX_MESSAGES=200
Verification
npm run build
npm test
DEEPSEEK_API_KEY="REPLACE_WITH_DEEPSEEK_KEY" npm run test:live
DEEPSEEK_MCP_AUTH_TOKEN="REPLACE_WITH_TOKEN" npm run test:remote
The live smoke test performs real DeepSeek requests for model listing, balance, non-thinking chat, thinking streaming chat with reasoning_content, FIM completion, and MCP tool calls.
Registry Identity
- MCP Registry name:
io.github.DMontgomery40/deepseek - npm package:
deepseek-mcp-server - OCI package:
docker.io/dmontgomery40/deepseek-mcp-server:0.5.0
Official References
- DeepSeek chat completions: https://api-docs.deepseek.com/api/create-chat-completion
- DeepSeek FIM completions: https://api-docs.deepseek.com/api/create-completion
- DeepSeek models: https://api-docs.deepseek.com/api/list-models
- DeepSeek balance: https://api-docs.deepseek.com/api/get-user-balance
- MCP specification: https://modelcontextprotocol.io/specification/2025-11-25
License
MIT
Recommended MCP Servers
0Latency MemoryPersistent memory layer for AI agents.
0nMCP — Universal AI API OrchestratorUniversal AI API Orchestrator — 1,554 tools, 96 services.
0xHumans Protocol MCPMCP for AI agents: financing, skills, lending on Base1k Patient Mcp1k patient MCP server
1trippulse1trip PULSE: 21-tool AI travel planner.
3D AI Agent Avatar3D AI Agent Avatar — render any GLB, give it a Solana wallet, a voice, and pump.fun powers.89How it compares
Official DeepSeek API MCP adapter, not a prompt library or RAG skill.
FAQ
Who is io.github.DMontgomery40/deepseek for?
Developers and developers who use DeepSeek as an LLM provider and want MCP-native chat, completion, model list, and balance tools in their agent IDE.
When should I use io.github.DMontgomery40/deepseek?
Use it in Build while wiring agents or scripts that must call DeepSeek, or when you need in-session model and balance visibility.
How do I add io.github.DMontgomery40/deepseek to my agent?
Install deepseek-mcp-server from npm, set DEEPSEEK_API_KEY, and register the stdio MCP server in Claude Code, Cursor, or another MCP-compatible client.