LLM-Client
vite + react + ts, simple LLM interface developed with user experience in mind.
Links
README
From the repo.
LLM Client(大模型小助手)🤖
An ultra-simplistic, completely local, npm-based Ollama web interface developed with user experience in mind. 极简 Ollama 本地网页端 UI

[!TIP]
This project is still under development, and is meant to serve as a early-stage side-project for Wanli-Go/Project-HILLM.
Features(功能)
- Ease of use: clone the repo, run
npm install, and runnpm run devto immediately use with your Ollama.- robust support for DeepSeek thinking models. (完全支持DeepSeek思考模式)
- Simplistic, Satisfactory UI: minimal configuration; strong focus on visual fluidness (with an option to reduce motion).
- Accessibility: use keyboard to navigate intuitively.
- Press
Tabto navigate between buttons. - Press
Ctrl + Lto open Settings.
- Press
- Lightweight: minimal dependencies (basic
vite+React+Typescript+ small packages likeprism.jsandprimeicons)- Custom-written markdown renderer.
- Language Support: Chinese first, English second.(中文优先)
Give a star or make PRs if you like it!
Installation(安装)
TL;DR: pull models -> clone repo -> check out .env -> run npm run dev
-
Install Ollama and Node.js, if you haven't already. (安装Ollama和Node.JS)
- Ollama 是一个在本地部署大模型的简单程序。
- Node.js 是一个 JavaScript 运行时。
-
Use Ollama to pull at least 1 DeepSeek model and 1 non-thinking model. (用Ollama获取至少一个DeepSeek模型和一个非思考模型)
- deepseek-r1
- 非思考模型,如 qwen2.5 (通义千问2.5)
-
Clone the repo, navigate to project's folder, and run
npm install. (克隆本仓库,在命令行上进入这个项目的目录,然后运行安装指令npm install) -
Configure
.env: (设置项目环境变量)
VITE_API_OLLAMA_URL:Ollama 对话接口(默认为http://localhost:11434/api/chat)VITE_THINKING_MODEL_NAME:Ollama官方模型名称(DeepSeek)VITE_GENERAL_MODEL_NAME:Ollama官方非思考模型名称
-
run
npm run dev. (在项目目录的命令行上运行命令启动程序) -
Happy Hacking!
Tips
- The project is still in her early stages.
- You can view the request body in browser's console every time you make a request.
- When you press the Regenerate button, it will use the input field's latest value, not last request's value.
Showcase


Collected info
- ★ 8 stars
- Language: TypeScript
- Source updated: 3/24/2025
Config for your environment
Replace {MCP_ENDPOINT_URL} with this MCP’s endpoint URL (from its repo or docs above). No API key — you connect directly.
Tool
OS
Config file: ~/.cursor/mcp.json
{
"mcpServers": {
"mcp-server": {
"url": "{MCP_ENDPOINT_URL}"
}
}
}Paste into mcpServers in the config file. Restart Cursor after saving.
If this MCP is also published on mcpchannel.ai, you can subscribe from Browse and use the gateway config there instead.