Canvas LMS MCP Server

# Amazon Q CLI Integration This document explains how to integrate the Canvas LMS MCP Server with Amazon Q CLI (qchat). ## Prerequisites - Amazon Q CLI installed - Canvas API token from your institution - Python 3.10+ with required dependencies ## Installation Steps ### 1. Clone and Install ```bash git clone https://github.com/sweeden-ttu/canvas-lms-mcp.git cd canvas-lms-mcp pip install -e . ``` ### 2. Configure Credentials ```bash cp .env.example .env # Edit .env with your Canvas API token and base URL ``` ### 3. Add MCP Server to Q CLI ```bash q mcp add --name canvas-lms --command python --args "/path/to/canvas-lms-mcp/server.py" ``` ### 4. Verify Installation ```bash q mcp list q mcp status --name canvas-lms ``` ## Usage Once configured, the Canvas MCP server provides these tools in Q CLI chat sessions: - `canvas_get_profile` - Get your Canvas user profile - `canvas_list_courses` - List enrolled courses - `canvas_get_todo` - Get to-do items - `canvas_get_assignments` - Get course assignments - `canvas_get_grades` - Get grades for a course - `canvas_get_announcements` - Get course announcements - `canvas_get_discussions` - Get discussion topics - `canvas_get_modules` - Get course modules - `canvas_get_upcoming_events` - Get calendar events ## Example Queries Ask Q CLI questions like: - "What assignments are due this week in my Canvas courses?" - "Show me announcements from all my classes" - "What's my current grade in course 12345?" - "List all my active courses" ## Troubleshooting ### Server Not Loading - Verify absolute path in MCP configuration - Check Canvas API token is valid - Ensure Python dependencies are installed ### 401 Unauthorized - Regenerate Canvas API token - Verify token has proper permissions ### 403 Forbidden - Some endpoints require instructor privileges - This is normal for student accounts ## Configuration Files - **Global**: `~/.aws/amazonq/mcp.json` - **Agent-specific**: `.amazonq/cli-agents/{agent-name}.json` For agent-specific configuration, add to agent JSON: ```json { "mcpServers": { "canvas-lms": { "command": "python", "args": ["/path/to/canvas-lms-mcp/server.py"] } } } ```