Coder (Remote Dev) MCP. Manage builds and workspace logs via chat.

Listen up. This server connects your AI agent straight into the backend of your remote development environment. You can manage everything from building pipelines to debugging agents without leaving your chat window.

Infrastructure and Workspace Control

You've got full control over who uses the system and where they work. You can set up a brand new corporate organization using create_organization, or view all existing ones with list_organizations. For managing personnel, you can add users via create_user and update their details using update_user_profile; you also manage roles by calling list_org_roles and listing members with list_org_members.

The system lets you create dedicated development workspaces with create_workspace, view all available ones with list_workspaces, or get specific metadata about a workspace instance using get_workspace. You can modify who has access by calling update_workspace_acl or checking the current rules with get_workspace_acl. When you need credentials, you generate user tokens via create_user_token and list all existing ones with list_user_tokens; for security, you can also create secrets using create_user_secret and view them with list_user_secrets.

To keep things clean, you can delete users with delete_user or manage the enterprise licenses by listing them with list_licenses, adding new ones via add_license, or removing old ones with delete_license.

Build and Deployment Management

You can monitor your entire deployment lifecycle. You check build versions and metadata using get_build_info, and you pull current metrics on deployment status, like successful builds or failures, with get_deployment_stats. If you need to know if the platform is outdated, run get_update_check. For managing workflows, you can list available templates through list_templates and check their historical versions using list_template_versions.

You initiate long-running background tasks for users using create_user_task, then manage them by listing all active jobs with list_tasks. To get the necessary connection details, you fetch the SSH config parameters with get_ssh_config or grab basic API info at get_api_root; you also retrieve deployment settings via get_deployment_config and check the current build requirements using get_prebuild_settings.

Debugging Agents and Workspaces

This is where the real power is. You can stream live logs from a specific workspace agent instance with get_agent_logs, or you can send accumulated log data from your local machine up to the server using send_agent_logs. For deep troubleshooting, you pull the last 10MiB of logs via get_agent_debug_logs, retrieve debug information from the Tailscale magicsock with get_agent_debug_magicsock, and check the startup manifest for the local agent by calling get_agent_debug_manifest.

You list all active containers and devcontainers in a workspace using get_agent_containers, and you set up network connectivity by creating port shares via create_workspace_port_share or listing existing ones with list_workspace_port_shares. The agent needs to connect, so you first get the necessary DERP details using get_agent_connection, then authenticate it with cloud provider credentials using auth_workspace_agent.

If a build is running too long, you stop it instantly with cancel_workspace_build. You can grab detailed info on one specific build run through get_workspace_build and pull all its parameters or full logs using get_workspace_build_params or get_workspace_build_logs.

AI Session Monitoring and Chat Control

You manage the AI interactions right here. You list active chat sessions with list_ai_sessions, or you start a brand new conversation history using create_chat. To send content to an ongoing discussion, you call create_chat_message; if you need context, you retrieve all messages from that session via get_chat_messages or get the associated thread details with get_ai_session_threads.

You can view every available AI model using list_ai_models, and list connected agents/clients through list_ai_clients. For usage tracking, you pull metrics on Daily Active Users (DAU) with get_insights_daus, calculate the activity duration per user account via get_insights_user_activity, or see template usage data using get_insights_templates.

System Auditing and Security

For compliance, you retrieve a paginated list of system audit logs with get_audit_logs. You verify if the user has permission for any action by running check_auth, and you manage external account links (like GitHub) using list_external_auth or starting the process through get_external_auth_device.