PPTAgent
An Agentic Framework for Reflective PowerPoint Generation
claude mcp add --transport stdio icip-cas-pptagent uvx pptagent \ --env OFFLINE_MODE="false" \ --env MINERU_API_URL="placeholder or local MinerU URL" \ --env PPTAGENT_CONFIG_PATH="path/to/config (if applicable)"
How to use
PPTAgent is an MCP server that integrates AI-assisted presentation generation into your workflow. It exposes a set of command-line tools via the uvx-based Package Management Interface, enabling you to onboard the service, configure your environment, and generate PPTX files from prompts. The available CLI commands include onboarding to set up credentials and preferences, generating presentations from natural language descriptions, viewing current configuration, and resetting configuration when needed. When used in offline mode, PPTAgent can operate with local assets while offering guidance on how to connect to external services when online capabilities are available. Tools like Deep Research Integration, Free-Form Visual Design, Autonomous Asset Creation, and Text-to-Image Generation are highlighted in the project notes, indicating that the server can orchestrate multiple sub-agents for data gathering, design proposals, and asset generation to produce a finished presentation.
How to install
Prerequisites:
- A system with Python and Node/npm tooling or a uv-based workflow as described by the project (uv/uvx installation steps are provided in the repo).
- Access to the internet to install required packages or a local offline setup if you plan to run in offline mode.
Install and configure PPTAgent MCP server:
- Install the uv toolchain (as shown in the Quick Start):
# Install uv (as per project instructions)
curl -LsSf https://astral.sh/uv/install.sh | sh
- Install the PPTAgent MCP tooling via uvx:
# On first run, onboard and configure PPTAgent
uvx pptagent onboard
- Verify configuration and generate a test PPT to ensure the server works:
uvx pptagent generate "Sample Title: Hello World" -o test.pptx
- If you are running in a Docker/Compose environment, follow the project’s docker or compose setup steps (noting that the MCP config you provide will map to the pptagent commands as described above). If you run locally, ensure dependencies are installed and that you can invoke the CLI tools directly as shown in the Quick Start section.
Optional offline setup (if you need offline capabilities): follow the repository guidance to enable offline mode, configure local endpoints, and set OFFLINE_MODE in the mcp.json or environment as needed.
Additional notes
Tips and considerations:
- Ensure you have all required API keys and services configured (MinerU, Tavily, LLM endpoints, etc.) if you intend to enable full capabilities.
- Offline mode is supported with limited capabilities; use offline_mode: true to avoid network-dependent tools.
- The MCP server appears to use a uvx-based workflow; the recommended commands include onboarding, generating presentations, and managing configuration. If you encounter environment variable issues, review PPTAgent-specific constants and the mcp.json example to align environment values.
- When running in a team or CI environment, consider pinning the exact uvx version to avoid breaking changes in tooling.
- If you need to reset the flow, the PPTAgent CLI provides a reset option to re-establish a clean configuration state.
Related MCP Servers
AgentChat
AgentChat 是一个基于 LLM 的智能体交流平台,内置默认 Agent 并支持用户自定义 Agent。通过多轮对话和任务协作,Agent 可以理解并协助完成复杂任务。项目集成 LangChain、Function Call、MCP 协议、RAG、Memory、Milvus 和 ElasticSearch 等技术,实现高效的知识检索与工具调用,使用 FastAPI 构建高性能后端服务。
python -client
支持查询主流agent框架技术文档的MCP server(支持stdio和sse两种传输协议), 支持 langchain、llama-index、autogen、agno、openai-agents-sdk、mcp-doc、camel-ai 和 crew-ai
ultimate_mcp_server
Comprehensive MCP server exposing dozens of capabilities to AI agents: multi-provider LLM delegation, browser automation, document processing, vector ops, and cognitive memory systems
MCP-PostgreSQL-Ops
🔍Professional MCP server for PostgreSQL operations & monitoring: 30+ extension-independent tools for performance analysis, table bloat detection, autovacuum monitoring, schema introspection, and database management. Supports PostgreSQL 12-17.
mcp-tool-kit
Agentic abstraction layer for building high precision vertical AI agents written in python for Model Context Protocol.
Common_Chronicle
Common Chronicle turns messy context into structured, sourced timelines.