spacecake

Product Introduction

1. Definition: Spacecake is an open-source desktop application functioning as an AI-augmented development environment (AI-IDE). It specifically integrates a Notion-style markdown editor with the Ghostty terminal emulator, designed to enhance workflows using Anthropic's Claude Code for AI agent development and spec-driven programming.
2. Core Value Proposition: It exists to bridge the gap between terminal-based AI agent execution and the need for a visual, efficient markdown editing interface. Spacecake provides real-time visibility into Claude's context window usage, operational costs ($), and agent task status, enabling developers to manage AI-driven coding workflows more effectively and cost-consciously within a unified local environment.

Main Features

1. Integrated Ghostty Terminal:
   How it works: Spacecake embeds the Ghostty terminal emulator directly into the application interface. This integration allows Claude Code running in the terminal to be contextually aware of the active file being edited in the markdown editor. Actions like edit <filepath> executed by Claude in the terminal automatically sync with the editor.
   Technologies: Ghostty (terminal emulator), custom IPC (Inter-Process Communication) for editor-terminal sync.
2. Real-Time Context & Cost Monitor:
   How it works: A persistent status bar displays critical metrics: the current context window utilization percentage (e.g., 🧠 67%) and the cumulative cost of the active Claude session in dollars (e.g., 💰 $4.20), dynamically updating as the session progresses. It specifies the Claude model version in use (e.g., Opus 4.5).
   Technologies: Anthropic API usage tracking, real-time data polling/streaming.
3. AI Agent Task Management Panel:
   How it works: Provides a live, filterable task board (pending, in-progress, completed) showing tasks generated and executed by Claude Code agents. Each task displays its title, status, and assigned agent owner (e.g., agent-1). Tasks automatically sync with agent activity in the terminal.
   Technologies: Task state parsing from terminal output/Claude interactions, state management for UI updates.
4. CodeMirror-Powered Markdown Editor:
   How it works: Offers a WYSIWYG (What You See Is What You Get) editing experience similar to Notion, specifically optimized for creating and editing technical markdown documents (CLAUDE.md, ship-faster.md). Supports rich code blocks, Mermaid diagrams for visual planning, and facilitates easy navigation between specifications and implementation code.
   Technologies: CodeMirror editor framework, custom extensions for WYSIWYG markdown and diagram rendering.

Problems Solved

1. Pain Point: Terminal inefficiency for markdown editing. Developers using Claude Code in terminals face significant friction when creating, editing, and visualizing complex markdown specs, plans, and documentation (CLAUDE.md files), hindering spec-driven development.
2. Target Audience:
   • Claude Pro Developers: Users leveraging Claude Code for AI agent-assisted software development.
   • Technical Writers/Planners: Individuals creating detailed technical specifications, research plans, and documentation using markdown within AI workflows.
   • AI Agent Builders: Developers building and managing teams of Claude agents for complex task execution.
3. Use Cases:
   • Spec-Driven Development: Writing CLAUDE.md files with clear instructions, stack definitions (e.g., "stack": "next.js"), and conventions while simultaneously running agents that execute against these specs.
   • Agent Workflow Monitoring: Tracking the progress of multiple AI agents (agent-1, agent-2) working on tasks like create database schema or build api endpoints in real-time.
   • Cost-Optimized Claude Usage: Preventing unexpected API costs by monitoring real-time expenditure (💰 $4.20) and context usage (🧠 67%) during extended development sessions with Claude Opus.
   • Visual Technical Planning: Using Mermaid diagrams within markdown to visually plan and explain complex systems or code structures to Claude.

Unique Advantages

1. Differentiation: Unlike generic IDEs or standalone terminals, Spacecake is purpose-built for Claude Code workflows. It uniquely combines a production-grade markdown editor (superior to basic terminal editors like Vim/Nano for markdown) with deep terminal integration and real-time Claude-specific analytics (context/cost). It contrasts with web-based Anthropic interfaces by being a local, open-source desktop app focused on developer productivity and agent management.
2. Key Innovation: The bidirectional sync between the visual editor and the Ghostty terminal, coupled with real-time Claude operational telemetry (context window, cost) presented within the UI. This creates a unified interface for AI-augmented development where specs, agent activity, terminal output, and cost metrics are visible and manageable in a single glance.

Frequently Asked Questions (FAQ)

1. Is Spacecake compatible with my operating system?
   Spacecake currently provides native builds for macOS (Apple Silicon and Intel) via .dmg files, and Linux (x86-64) via .deb packages and .AppImage formats. Support for other platforms (like Windows) is planned ("other platforms coming soon").
2. How does Spacecake differ from using Claude in the web interface or a standard terminal?
   Spacecake integrates a professional markdown editor directly with the terminal running Claude Code, providing real-time cost/context tracking and visual agent task management – features absent in the standard web chat or isolated terminal environments. It's designed for complex, multi-step AI agent development.
3. What are the benefits of Spacecake being open-source?
   Being open-source (hosted on GitHub) ensures transparency, allows community inspection for security, enables customization to specific workflows, and facilitates community contributions for features and bug fixes.
4. How does the task management panel work with Claude agents?
   As Claude Code agents execute commands (like edit, run, build), they generate and update tasks within the Spacecake interface. The panel automatically syncs with this activity, displaying task status (pending, in-progress, completed), title, and owner, providing a live overview of the agent workflow.
5. How accurate is the real-time cost tracking?
   The cost tracking (💰 $4.20) estimates expenditure based on Anthropic's published pricing for the active model (e.g., Claude Opus) and the volume of tokens processed (input + output) within the current session. It provides a reliable real-time estimate for budget management during development.