# crystl Brand Primary

## Brand Essence

**crystl is the multitasking terminal for AI coding agents.** It gives developers a single, purposeful macOS workspace to run, organize, and oversee multiple AI-assisted coding sessions—without being locked into one agent, model, or workflow.

At its core, crystl turns the terminal from a collection of disconnected windows into a clear operating environment for agentic software development. Projects are organized into **gems**; isolated parallel sessions run in **shards** and separate worktrees; developers retain visibility and control over what each agent is doing.

### Core Identity
crystl is an agent-agnostic, developer-controlled terminal built for the reality of modern coding: multiple projects, multiple parallel tasks, and multiple AI agents working at once. It supports tools such as **Claude Code, Codex, and Antigravity CLI** while preserving the user’s own rules, tools, and preferred models.

### Mission
Enable developers to productively orchestrate AI coding agents in parallel—without sacrificing clarity, isolation, ownership, or control.

### Vision
A development workflow where AI agents are flexible collaborators rather than opaque tools: easy to deploy across tasks, easy to monitor, and governed by the developer’s standards.

### Key Value Propositions
- **Agent and model freedom:** Use any compatible coding agent and model rather than adopting a closed ecosystem.
- **Parallel work without chaos:** Run multiple sessions simultaneously in isolated worktrees.
- **Structured organization:** Group projects into gems and sessions into manageable shards.
- **Always-on visibility:** See what every agent is doing from one cohesive terminal environment.
- **Developer control:** Keep workflows aligned with “your rules,” not a platform’s prescribed process.
- **Native, considered experience:** A macOS-first terminal with a distinctive frosted-glass interface that makes complex work feel calm and legible.

## Market Positioning

crystl sits at the intersection of developer tooling, terminal productivity, and AI-agent orchestration. It is designed for developers and technical teams who have moved beyond using one coding assistant in one terminal tab and now need a reliable way to manage concurrent agent-driven work.

### Target Audience
Primary audiences include:
- Independent developers and power users running AI coding agents across several tasks or repositories
- Agency owners and small software teams coordinating parallel client and product work
- Technical leaders and early adopters building agent-native engineering workflows
- macOS developers who value terminal-native workflows, visual clarity, and direct control

### Competitive Differentiation
Unlike a conventional terminal, crystl is purpose-built for multitasking with AI coding agents. Unlike a single-agent IDE or proprietary agent platform, it is **agent- and model-agnostic**. Its defining advantage is combining parallel execution, isolated worktrees, project organization, and real-time oversight in one native environment.

crystl does not position AI as an autonomous replacement for developers. It positions AI agents as work that developers actively direct, structure, and supervise.

### Brand Personality and Tone
The crystl voice is confident, concise, technical, and developer-respectful. It should feel modern and visually refined without becoming vague or overly futuristic. Favor direct language, concrete workflow benefits, and user agency.

Key tonal qualities:
- **Empowering:** The developer remains in charge.
- **Clear:** Complexity is made visible and manageable.
- **Open:** Any agent. Any model.
- **Sophisticated:** Premium macOS craft paired with practical utility.
- **Pragmatic:** Built from real experience using agents in production work.

## Key Messaging

### Primary Brand Message
**Any agent. Any model. Your rules.**

crystl provides the workspace to run AI coding agents in parallel while keeping projects organized, sessions isolated, and every agent visible.

### Supporting Narratives
- AI coding is becoming inherently multitasking; the terminal should evolve to support it.
- Parallel agents are more useful when their work is separated into clean, isolated worktrees.
- Developers should not have to choose between speed and oversight.
- The best agent workflow is flexible: use the tools and models that fit the task.
- A polished interface can reduce operational noise without abstracting away developer control.

### Proof Points and Credibility Signals
- Built specifically for **macOS** and AI coding-agent workflows
- Supports leading agent tools including **Claude Code, Codex, and Antigravity CLI**
- Organizes work through named concepts—**gems** for projects and **shards** for isolated sessions
- Enables concurrent agent activity through separate worktrees
- Originated from firsthand experience managing multiple AI agents in an agency context