Coder vs Eclipse Che

Coder is a self-hosted platform for AI coding agents and cloud development environments running on infrastructure controlled by the customer. Workspaces are defined with Terraform and connect through a secure WireGuard tunnel. Developers can access them through VS Code, JetBrains, Cursor, a web terminal, remote desktop or SSH. Idle workspaces can stop automatically and restart when needed. Coder Agents lets developers delegate development work and research through chat or an API, provisioning a workspace to carry out tasks. Supported model providers include Anthropic, OpenAI, Google, Bedrock and self-hosted endpoints. The agent loop runs in Coder’s control plane, keeping LLM credentials out of workspaces. Coder documents integrations with GitHub, GitLab, Bitbucket and JFrog Artifactory. The Community plan costs nothing and includes unlimited self-hosted workspaces, one platform integration and up to five active agents. Community support is through GitHub and Discord. The CLI is available for Linux, macOS and Windows, alongside API, web and self-hosted platforms.

Eclipse Che is an open-source development environment that runs IDEs in Kubernetes. Users create workspaces from a Git repository or sample, then access them online through a browser. Visual Studio Code and JetBrains IDEs are available. Workspace definitions use Devfiles, YAML files that can be versioned with application source code, and a devcontainer configuration is listed. GitHub, GitLab, and Bitbucket identities can connect through OAuth. Teams can share a development setup by URL, and Che can run on public or private Kubernetes clusters, including self-hosted deployments. Access uses OIDC authentication and Kubernetes RBAC authorization. Enterprise support includes air-gapped networks, TLS certificates, proxies, and FIPS. Code is cloned to persistent storage by default, but other file changes disappear when a workspace stops unless stored on a persistent volume. Che is free, with no stated usage limits.