At a Glance:
Coder is a self-hosted platform for cloud development environments and AI coding agents, supporting Terraform-defined workspaces, automated idle resource shutdown, and a native AI agent that executes on your infrastructure without exposing LLM credentials in workspaces.
Overview:
Coder is a self-hosted platform that provisions cloud development environments and runs AI coding agents. It allows teams to define development workspaces as code through Terraform, supporting infrastructure types such as EC2 VMs, Kubernetes pods, and Docker containers. A built-in automatic shutdown mechanism stops idle resources to control costs. Coder also includes Coder Agents, a native AI coding agent whose execution loop resides in the control plane on your own infrastructure, preventing API keys from appearing inside individual workspaces. The platform centralizes model governance, cost tracking, and audit logging while supporting bring-your-own-model configurations across providers like Anthropic, OpenAI, Google, Bedrock, and self-hosted models.
Key Decision Points:
Infrastructure as code: Workspaces are defined with Terraform, making environment provisioning repeatable and version-controlled.
Self-hosted control plane: All components run on your own infrastructure, including the AI agent's execution loop.
AI credential isolation: AI agent actions are authenticated through the control plane, so no LLM API keys reside in workspaces.
Idle resource management: Workspaces automatically shut down when not in use, which directly impacts cloud costs.
Core Features:
Terraform-defined workspaces: Provision cloud development environments as EC2 VMs, Kubernetes pods, or Docker containers using Terraform templates.
Idle shutdown automation: Automatically stop idle resources to reduce infrastructure spending.
Native AI coding agent: Coder Agents executes coding tasks with its loop running in the control plane, not in the workspace.
Model governance: Choose from Anthropic, OpenAI, Google, Bedrock, or self-hosted models with centralized control and audit logging.
IDE integrations: Open workspaces through both VS Code and JetBrains Gateway with one-click plugins.
Use Cases:
Platform engineers provisioning standardized, on-demand development environments on cloud infrastructure using Terraform.
Engineering teams onboarding developers with pre-configured workspaces instead of manual local setup processes.
Organizations adopting AI coding assistants who need centralized model governance, cost visibility, and credential isolation from development environments.
Open-Source Alternative Value:
Coder is a self-hosted platform available as an open-source project, offering organizations a way to manage cloud development environments and AI coding agents on their own infrastructure. Workspaces are built from version-controlled Terraform definitions, and idle resources are automatically shut down. For AI-assisted development, the native agent architecture keeps all LLM interactions in the control plane, supporting centralized model selection, cost tracking, and audit logging without placing API keys inside workspaces.
