Introduction

Shellarium is a desktop application that lets you manage multiple AI coding agents in parallel. Instead of writing code yourself, you describe tasks and let AI agents do the work — then review, approve, and merge their changes.

Current version: v0.7.89.

Why Shellarium?

Traditional IDEs are built around a code editor — you write code. Shellarium is built around a different premise: AI agents write code, you review and orchestrate.

Key Capabilities

• Parallel execution — run dozens of agents simultaneously, each in an isolated git worktree
• Multi-provider — use Claude Code natively, or install Codex CLI, Aider, and other agents via the ACP registry
• Built-in review — diff viewer with syntax highlighting, merge workflow, GitHub integration
• Permission control — five modes from full manual approval to fully autonomous, with path-based rules and dangerous command detection
• Git-native — every agent works in its own worktree, zero conflicts between agents
• Automations — schedule recurring tasks with cron, trigger on GitHub events, or run manually
• Rich attachments — paste images, capture screenshots, attach files to agent prompts