Happy slice of mcp pizzaMCP.pizza
project-golem

project-golem

MCP.Pizza Chef: Arvincreator

This aims to be a always-on hub rather than a chat window: it keeps long-term memory so context survives between sessions, coordinates several agents on one job, and runs work on a schedule without prompting. It can connect to hosted models or ones running on your own machine, and comes with a dashboard for watching what is happening. Setup is substantial and needs a recent Node.

Coding
Communication
Notes

Use This MCP server To

Keep context between separate sessions Run a job on a schedule automatically Coordinate several agents on one task Watch what my agents are doing Use a model running on my own machine Pick up a task where I left off

README

Project Golem Logo

Project Golem

你的 AI 操作中樞：可對話、可行動、可記憶、可長期協作。

version node backend dashboard

English README · 貢獻指南 · 開發代理指南

Project Golem 是什麼

Project Golem 不是單純聊天機器人，而是一套可運行的 AI 工作系統。
它把「語言模型能力」和「實際操作能力」整合成同一個執行核心，讓你可以在一個 Dashboard 裡完成：

  • 即時對話與任務協作
  • 技能掛載與工具擴充（MCP / Skills）
  • 長期記憶管理與檢索
  • 多代理討論與共識彙整
  • RPG 劇情互動、股市看板、協作日曆、羈絆日記
  • 系統安全、設定、更新與維運

你可以使用：

  • Gemini Web（Browser-in-the-Loop，透過 Playwright）
  • Ollama（本機/私有模型）
  • LM Studio（OpenAI-compatible local API）

為什麼它有價值

project-golem FAQ

Can I use this to keep context between sessions?
Yes — long-term memory is one of its central features.
Can it run things on a schedule?
Yes, there is a scheduler for work that should happen without prompting.
Can I use a model on my own machine?
Yes — locally run models are supported alongside hosted ones.
Is there a way to see what it's doing?
Yes, it includes a dashboard for watching activity.
How hard is it to set up?
Substantial — it is a multi-part system needing a recent Node and configuration.