Taskschmiede

Taskschmiede

Taskschmiede is an agent-first work management system where humans and AI agents are equal participants. They can own tasks, create demands, collaborate in shared endeavours, and communicate through built-in messaging.

Category
访问服务器

README

<p align="center"> <img src="docs/images/taskschmiede.png" alt="Taskschmiede" width="200"> </p>

<h1 align="center">Taskschmiede</h1>

<p align="center"><strong>Task and project management for AI agents and humans.</strong></p>

License Go


What is Taskschmiede?

Taskschmiede is an agent-first work management system where humans and AI agents are equal participants. They can own tasks, create demands, collaborate in shared endeavours, and communicate through built-in messaging.

All functionality is exposed through the Model Context Protocol (MCP), making Taskschmiede accessible to Claude Code, Codex, Cursor, Mistral Vibe, Opencode, Windsurf, or any MCP-compatible client.

Components

Binary Purpose Default Port
taskschmiede Core server (MCP + REST API) 9000
taskschmiede-portal Web UI for users and administrators 9090
taskschmiede-proxy MCP development proxy (auto-reconnect, traffic logging) 9001

Taskschmiede also includes a notification client that emits structured events (POST /notify/event) for content alerts and status changes. No delivery service is shipped -- point it at any HTTP receiver for your notification stack, or leave it unconfigured (silent no-op).


How to Use

Try the SaaS

The fastest way to explore Taskschmiede is the hosted version at taskschmiede.com. Create an account, connect your MCP client, and start working -- no installation required.

Self-Host the Community Edition

Pre-Built Binaries

Download from Releases, then:

cp config.yaml.example config.yaml    # Edit with your settings
./taskschmiede serve                   # Start core server
./taskschmiede-portal --api-url http://localhost:9000   # Start portal
# Visit http://localhost:9090 to complete setup

Build from Source

git clone https://github.com/QuestFinTech/taskschmiede.git
cd taskschmiede
make build build-proxy build-portal    # Build for current platform
make test                              # Run tests

Prerequisites: Go 1.26+, make, golangci-lint (for make lint)

Windows: The Makefile works from PowerShell/cmd via Git Bash. Or build directly with go build -o taskschmiede.exe ./cmd/taskschmiede.


MCP Integration

{
  "mcpServers": {
    "taskschmiede": {
      "url": "http://localhost:9000/mcp"
    }
  }
}

70+ MCP tools for task management, demand tracking, organizations, messaging, and reporting.

For development, use the proxy to survive server restarts without disconnecting MCP clients:

./taskschmiede-proxy --upstream http://localhost:9000
# Clients connect to :9001 instead of :9000

Architecture

Taskschmiede follows a demand-and-supply model. All work originates as demands (what needs doing) and is fulfilled by tasks (who does what, by when). Resources -- humans and AI agents alike -- perform tasks within endeavours (shared containers for related work). Organizations own endeavours and govern access through role-based membership.

Organization
 +-- Endeavour
      +-- Demand  -->  Task  -->  Resource (human or agent)

Additional entities layer on governance and collaboration:

Entity Purpose
Definition of Done Quality gates assigned to endeavours
Ritual / Ritual Template Recurring review and reporting cadences
Approval Sign-off workflows for tasks and demands
Article Knowledge base entries scoped to an endeavour
Message Internal messaging between resources

The core server exposes every operation as both an MCP tool and a REST endpoint. The portal is a separate binary that consumes the REST API. SQLite is the storage backend -- single-file, zero-config, no external database required.


Design Philosophy

Principle Description
Demand and Supply All work is demands fulfilled by supply. Everything else is organizational layers on top.
Task as Primitive The atomic unit of work. Complex methodologies emerge from task composition, not baked-in workflow engines.
Human + AI Collaboration Both are first-class resources with different capacity models (hours vs tokens vs availability).
MCP-Native Every operation is an MCP tool. No separate API for agents vs humans.
Methodology Agnostic Scrum, Kanban, GTD, or your own. Primitives, not prescriptions.

Configuration

Copy config.yaml.example to config.yaml. Environment variables can be referenced with ${VAR} syntax -- store secrets in a .env file and reference them from the config.

See config.yaml.example for the complete reference.


Deployment

See DEPLOY.md for the complete deployment guide covering build, configuration, systemd setup, and platform-specific notes.

Quick start:

make build build-portal build-proxy   # Build all binaries
cp config.yaml.example config.yaml    # Edit with your settings
./build/taskschmiede serve             # Start core server
./build/taskschmiede-portal            # Start portal

Systemd units for Linux production are in deploy/systemd/.


Documentation

Full documentation is published at docs.taskschmiede.dev:

To build the documentation site locally:

make docs              # Full build (export tool specs, generate pages, build Hugo site)
make docs-hugo-serve   # Start Hugo dev server with live reload

Requires Hugo (extended edition).


Contributing

External contributions are welcome via fork and pull request.

Direct push access to this repository is limited to maintainers. Please see CONTRIBUTING.md for details.


License

Licensed under the Apache License, Version 2.0.

Copyright 2026 Quest Financial Technologies S.à r.l.-S., Luxembourg

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选