What is Conan?

Conan is a native macOS app that wraps Claude Code in a live HUD — every prompt, tool call, skill, and token, surfaced as it happens. It runs on macOS 13+ with Apple silicon and costs $29 once, with no subscription.

Who it's for

• Developers using Claude Code — who want real-time visibility into every command, tool call, and context decision their AI agent makes
• Power users of AI coding tools — who need to understand token burn, context window usage, and skill execution to optimize their workflows
• Observability-minded builders — who prefer a live dashboard over opaque terminal output when working with agentic AI

Key features

Timeline

Every command, edit, and tool call streams onto a living timeline — the whole session, as it happens. You see prompts, skill firings, tool executions, and agent stops in chronological order, with timestamps and token counts for each event.

Context

Watch the context window fill in real time — system instructions, tools, memory, skills, and messages — and see exactly how many tokens each component consumes. A live breakdown shows what's eating your budget, from system overhead to skill payloads.

Pulse

A live throughput meter shows spikes when Claude is working and calm when it waits. You can feel the session's heartbeat — useful for spotting stalls, heavy tool calls, or unexpected pauses.

Skills & MCP

Every skill and MCP server is surfaced and observable. You see which skills fired, when they last ran, and where they live on disk. No more guessing what's in play — the HUD shows you the full arsenal.

What stands out

"Every prompt, tool call, skill, and token, surfaced as it happens."

Conan transforms Claude Code from a black-box terminal into a transparent, observable campaign. Instead of wondering what your AI agent is doing or why it burned through tokens, you get a live HUD that shows the entire decision chain — from skill selection to context pressure to tool execution. It's the difference between watching a battle from a hilltop versus being blind in the fog.

Worth checking out if…

You use Claude Code regularly and want to understand exactly what your agent is doing, how many tokens it's burning, and which skills it's reaching for. The $29 one-time price makes it a no-brainer for anyone who values visibility into their AI workflow — especially if you're optimizing for token efficiency or debugging complex multi-step sessions.