mcpappwrite

A minimal Python MCP Todo server backed by Appwrite Cloud.

The project is intentionally small: ChatGPT/MCP clients call Todo tools, and the server stores the data in an Appwrite TablesDB table.

Architecture

MCP client / ChatGPT
        |
        v
   Python MCP server
        |
        v
   Appwrite TablesDB
        |
        v
      tasks

Requirements

• Python 3.10+
• An Appwrite Cloud project
• An Appwrite server API key with row read/write access

Appwrite setup

Create a database and a table named tasks (the IDs can also be custom):

• Database ID: todo_db
• Table ID: tasks

Create these columns:

Column	Type	Required	Notes
title	Varchar	Yes	Task title
description	Text	No	Details
dueDate	Datetime	No	ISO-8601 datetime
priority	Integer	No	1 high, 2 medium, 3 low
completed	Boolean	No	Default false
category	Varchar	No	Example: Career
tags	Varchar	No	Comma-separated tags

For the server API key, grant only the database/table/row scopes required for this project. Do not put the API key in source code.

Local setup

Windows PowerShell

python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
Copy-Item .env.example .env

macOS/Linux

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env

Edit .env with your Appwrite values.

APPWRITE_ENDPOINT=https://<REGION>.cloud.appwrite.io/v1
APPWRITE_PROJECT_ID=your_project_id
APPWRITE_API_KEY=your_server_api_key
APPWRITE_DATABASE_ID=todo_db
APPWRITE_TABLE_ID=tasks

Run

The server uses MCP Streamable HTTP so it can later be deployed as a remote MCP server.

python app.py

The MCP endpoint is normally:

http://localhost:8000/mcp

For a quick local tool test, use the MCP Inspector or another MCP client.

Available tools

• add_task - create a task
• list_tasks - list tasks, optionally filtering completed status/category
• get_task - retrieve one task by row ID
• update_task - update selected task fields
• complete_task - mark a task complete
• delete_task - delete a task

Security

.env is ignored by Git. Never commit an Appwrite API key. For production, put the secret in the hosting platform's secret/environment-variable system.

Future steps

1. Test Appwrite connection locally.
2. Test every MCP tool locally.
3. Add authentication for a remote deployment.
4. Deploy the MCP server on a free-tier host.
5. Connect the remote MCP server to a compatible MCP client.