PromptVault MCP Server
PromptVault
作为 MCP 服务器的开源提示词版本管理、评估和管理工具。
什么是 PromptVault?
PromptVault 是一个面向 AI 工程师和提示词工程师的自托管工具,用于:
对提示词进行版本管理,像管理代码一样,支持不可变版本和回滚
评估提示词:使用 LLM 提供商(OpenAI、Anthropic、Ollama)基于数据集进行评估
访问所有内容:通过 MCP 服务器、CLI 或 REST API
Related MCP server: PromptingBox MCP Server
快速开始
安装(无需克隆)
pip install pvlt或者使用 uv/pipx:
uvx --from git+https://github.com/KrishBnsl/promptVault.git promptctl --help或者使用 Docker:
docker run -p 8000:8000 -e GEMINI_API_KEY=your-key ghcr.io/krishbnsl/pvlt从源码安装
git clone https://github.com/KrishBnsl/promptVault.git
cd promptVault
uv sync
cp .env.example .env
# Edit .env with your API keys创建你的第一个提示词
promptctl prompt create summarize \
--content "Summarize the following article in {tone} style:\n\n{article}" \
--description "Summarizes an article" \
--variables '{"tone": "concise", "article": "string"}' \
--model-config '{"provider": "openai", "model": "gpt-4o-mini", "temperature": 0.2}' \
--tags "summarization,content"创建数据集
echo '{"input": {"article": "AI is transforming..."}, "expected_output": "AI transforms..."}' > dataset.jsonl
promptctl dataset create article-summaries --file dataset.jsonl运行评估
promptctl eval run summarize --dataset article-summaries与 MCP(Claude Desktop)一起使用
添加到你的 Claude Desktop 配置中:
{
"mcpServers": {
"pvlt": {
"command": "promptctl",
"args": ["serve"]
}
}
}CLI 参考
promptctl prompt create <name> --content <text> [options]
promptctl prompt list [--tags TAGS] [--limit N]
promptctl prompt show <name> [--version N]
promptctl prompt versions <name>
promptctl prompt diff <name> <version-a> <version-b>
promptctl prompt rollback <name> --version N
promptctl dataset create <name> --file <jsonl-or-json>
promptctl dataset list
promptctl eval run <prompt-name> --dataset <dataset-name>
promptctl eval report <evaluation-id>
promptctl serve [--stdio|--http] [--port 8000]
promptctl web [--port 8080]MCP 工具
工具 | 描述 |
| 创建一个带有初始版本的新提示词 |
| 获取一个提示词版本 |
| 列出所有提示词 |
| 列出一个提示词的所有版本 |
| 比较两个提示词版本的差异 |
| 回滚到之前的版本 |
| 创建数据集 |
| 列出数据集 |
| 获取数据集及其条目 |
| 运行评估 |
| 检查评估状态 |
| 获取评估报告 |
| 比较两次评估 |
REST API
基础 URL: http://localhost:8000/api
# Create a prompt
curl -X POST http://localhost:8000/api/prompts \
-H "Content-Type: application/json" \
-d '{"name": "test", "content": "Hello {name}"}'
# List prompts
curl http://localhost:8000/api/prompts
# Get a prompt
curl http://localhost:8000/api/prompts/test配置
将 .env.example 复制为 .env 并进行配置:
PROMPTVAULT_DB_PATH=./promptvault.db
PROMPTVAULT_DEFAULT_PROVIDER=openai
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
GEMINI_API_KEY=...
OLLAMA_BASE_URL=http://localhost:11434架构
┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ MCP Client │ │ CLI │ │ REST Client │
│ (Claude etc) │ │ (promptctl) │ │ (curl, apps) │
└──────┬──────┘ └──────┬──────┘ └──────┬──────┘
│ │ │
▼ ▼ ▼
┌─────────────────────────────────────────────────┐
│ PromptVault Core │
│ ┌─────────────┐ ┌──────────────┐ ┌──────────┐ │
│ │ Versioning │ │ Evaluation │ │ Storage │ │
│ │ Engine │ │ Engine │ │ Layer │ │
│ └─────────────┘ └──────────────┘ └──────────┘ │
│ ┌──────────────────────────────────────────┐ │
│ │ SQLite Database (local) │ │
│ └──────────────────────────────────────────┘ │
└─────────────────────────────────────────────────┘文档
📖 完整文档: https://krishbnsl.github.io/promptVault/
许可证
MIT 许可证 - 详见 LICENSE。
This server cannot be deployed
Maintenance
Related MCP Connectors
- PromptOTOAuthcom.promptot
Manage, version, and publish LLM prompts with blocks, variables, and evaluations.
Generate contextual prompts and reusable agent skills, evaluate prompts with the 16-dimension Prompt Score, and manage saved work in PromptDrive. Twelve MCP tools also provide authorized access to private Memory for source-grounded answers. Connect over Streamable HTTP using OAuth 2.1 and PKCE. Generation consumes account quota and automatically saves successful results; Memory access follows account permissions and plan limits.
Self-hosted AI prompt library: prompts, collections, tags, teams, chains. 29 MCP tools for agents.
Professional prompt library over remote MCP: 13 verticals, free discovery scope.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables management and invocation of AWS Bedrock managed prompts with support for variable substitution, streaming responses, batch processing, and multiple AI models including Claude, Titan, Llama, and Mistral.MIT
- AlicenseAqualityDmaintenanceAllows users to save, organize, and manage AI prompts, folders, and tags directly within MCP-compatible tools like Claude and Cursor. It supports version history tracking, prompt searching, and the ability to save public templates to a personal collection.2011 npmMIT
- AlicenseNot gradedqualityCmaintenanceProvides prompt management, resource management, and tool call capabilities (content analysis, article analysis) based on the MCP protocol.3MIT
- AlicenseAqualityDmaintenanceEnables storage, versioning, and retrieval of prompts with rich metadata via MCP tools, supporting stdio and SSE transports.38 npm3ISC