A collection of reusable Terraform modules for extending Coder workspace functionality. These modules provide easy integration of popular development tools, applications, and services into your Coder workspace templates.

Integrate Anthropic's Claude Code AI assistant directly into your workspace.

Features:

• Web-based Claude Code interface with task reporting
• Optional CLI access through AgentAPI
• Automatic Node.js/npm installation via NVM
• Configurable installation scripts (pre/post-install)
• AgentAPI integration for enhanced AI capabilities

Usage:

module "claude_code" {
  source   = "git::https://github.com/AI-Riksarkivet/coder-modules.git//claude-code?ref=main"
  agent_id = coder_agent.main.id

  # Optional configuration
  folder                    = "/home/coder"
  claude_code_version      = "latest"
  experiment_report_tasks  = true
  install_agentapi         = true
}

High-performance, containerized VNC server for remote desktop access.

Features:

• Support for multiple desktop environments (XFCE, KDE, GNOME, LXDE, LXQT)
• Configurable ports and versions
• Built-in health checking
• Web-based access through browser

Usage:

module "kasmvnc" {
  source              = "git::https://github.com/AI-Riksarkivet/coder-modules.git//kasmvnc?ref=main"
  agent_id            = coder_agent.main.id
  desktop_environment = "xfce"
  port                = 6800
}

Visual framework for building multi-agent and RAG applications.

Features:

• Drag-and-drop interface for AI workflows
• Configurable base paths and ports
• Health monitoring and logging
• Subdomain support

Usage:

module "langflow" {
  source    = "git::https://github.com/AI-Riksarkivet/coder-modules.git//langflow?ref=main"
  agent_id  = coder_agent.main.id
  port      = 7860
  subdomain = true
}

Reactive Python notebooks that are reproducible, git-friendly, and deployable as scripts or apps.

Features:

• Interactive Python notebook environment
• Git-friendly notebook format
• Real-time reactivity and execution
• Health monitoring with configurable paths

Usage:

module "marimo" {
  source   = "git::https://github.com/AI-Riksarkivet/coder-modules.git//marimo?ref=main"
  agent_id = coder_agent.main.id
  port     = 2818
}

Open-source toolkit for building high-quality datasets and computer vision models.

Features:

• Dataset visualization and exploration
• Model evaluation and analysis
• Label and annotation management
• Automatic setup and configuration

Usage:

module "fiftyone" {
  source   = "git::https://github.com/AI-Riksarkivet/coder-modules.git//fiftyone?ref=main"
  agent_id = coder_agent.main.id
  port     = 5151
}

Command-line utility for sending notifications to Slack channels.

Features:

• Integration with Coder's external auth providers
• Customizable message templates
• Default channel configuration
• Command completion notifications

Usage:

module "slackme" {
  source           = "git::https://github.com/AI-Riksarkivet/coder-modules.git//slackme?ref=main"
  agent_id         = coder_agent.main.id
  auth_provider_id = coder_external_auth.slack.id
  default_channel  = "#dev-notifications"
}

Each module follows consistent patterns:

module-name/
├── main.tf              # Main Terraform configuration
├── run.sh              # Installation/startup script (if applicable)
├── scripts/            # Additional helper scripts
└── README.md           # Module-specific documentation

All modules require:

• agent_id - The ID of the Coder agent to install the module on

Most modules support:

• port - Port number for web applications (module-specific defaults)
• subdomain - Enable subdomain routing for web apps
• share - Access level (owner, authenticated, public)
• order - UI display order
• group - Logical grouping in the UI

1. Add Module to Template: Include the module in your Coder workspace template:

terraform {
  required_providers {
    coder = {
      source  = "coder/coder"
      version = "~> 2.0"
    }
  }
}

resource "coder_agent" "main" {
  arch = "amd64"
  os   = "linux"
  # ... agent configuration
}

# Add any module
module "claude_code" {
  source   = "git::https://github.com/AI-Riksarkivet/coder-modules.git//claude-code?ref=main"
  agent_id = coder_agent.main.id
}

2. Apply Template: Deploy the template to your Coder instance
3. Create Workspace: Launch a new workspace with the integrated modules

1. Create a new directory with the module name
2. Add main.tf with Terraform configuration
3. Include installation scripts if needed
4. Add health checks for web applications
5. Update this README with module documentation

Test modules by:

1. Creating a test workspace template
2. Including the module with various configurations
3. Verifying application startup and functionality
4. Testing health checks and UI integration

Web-based modules (langflow, marimo, fiftyone, kasmvnc) include:

• Health Checks: Automatic monitoring of application availability
• Subdomain Support: Optional subdomain routing for cleaner URLs
• Access Control: Configurable sharing permissions
• Logging: Structured logging to specified paths

Utility modules (claude-code, slackme) provide:

• Installation Scripts: Automated setup and configuration
• External Integrations: Authentication and service connections
• Custom Commands: CLI tools and shortcuts

1. Fork Repository: Create your own fork
2. Create Module: Follow the established patterns
3. Test Thoroughly: Verify functionality across environments
4. Update Documentation: Include comprehensive usage examples
5. Submit PR: Create pull request with detailed description

• Coder Documentation: Official Coder platform documentation
• Terraform Coder Provider: Provider documentation
• Module Development Guide: Terraform module best practices

For issues and questions:

• Module Issues: Create GitHub issues with detailed descriptions
• Coder Platform: Reference official Coder documentation and support
• Integration Help: Check individual module README files for specific guidance

This project is licensed under the MIT License. See the LICENSE file for details.