业务系统MCP服务中枢

这是一个私有MCP中枢,通过独立配置的Freshdesk与Trello服务,将ChatGPT安全对接到业务系统和运营数据,适用于客服与项目管理场景。

  • by Robert-Velhorst
  • Python
  • 本地部署
  • 更新于
在 GitHub 查看 Noodzakelijk-Online/040---Freshdesk-Trello-MCP-connectors-to-ChatGPT-5

使用说明

Connected Services MCP Hub

This private MCP hub connects ChatGPT to business systems and operational data. Each service is independently configured so credentials and access boundaries stay scoped to the external system.

Services

Service Directory Default port Required environment variables
Freshdesk freshdesk/ 8001 FRESHDESK_DOMAIN, FRESHDESK_API_KEY
Trello trello/ 8000 TRELLO_API_KEY, TRELLO_TOKEN

The Trello service is the canonical, full-featured implementation. It consolidates the former lightweight connector and enhanced Trello server in one location. The lightweight overview, search, fetch, and move_card tool names remain available as compatibility aliases; use the typed Trello tools for all new integrations.

Run Freshdesk

cd freshdesk
python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
Copy-Item .env.example .env
# Fill in the service credentials in .env, then:
python main.py

Run Trello

The Trello server includes board, list, card, checklist, collaboration, analytics, export, and automation tools. It is derived from m0xai/trello-mcp-server, retained under its Apache-2.0 license, and incorporates upstream PRs #11, #20, #21, #23, and #27 plus the current upstream checkitem fix.

cd trello
Copy-Item .env.example .env
# Set TRELLO_API_KEY and TRELLO_TOKEN in .env.
# For a ChatGPT read/fetch connector, also set:
# TRELLO_READ_ONLY=true
# For a network MCP connection, set:
# USE_CLAUDE_APP=false
uv sync --locked --python 3.15
uv run --python 3.15 python main.py

TRELLO_READ_ONLY=true is a server-side access boundary: mutation tools are not registered or exposed to the client. The server uses Streamable HTTP in network mode; connect your MCP client to the externally exposed endpoint (normally /mcp). Do not commit .env files, generated logs, board exports, or Trello credentials.

The Trello lockfile supports Python 3.12 and 3.13; its pinned pydantic-core does not currently build on Python 3.14.