jebi

Product Introduction

1. Definition: jebi is a native macOS terminal emulator and command-line interface (CLI) supercharged with a built-in, on-device AI assistant. It functions as a developer's primary workspace, combining a standard shell environment with integrated machine learning models for command suggestion, error analysis, and interactive chat.
2. Core Value Proposition: The product exists to eliminate friction in the terminal workflow by providing context-aware AI assistance without cloud dependency. Its core proposition is delivering a "terminal that thinks with you," offering local AI help for debugging, explanation, and command planning with a strong emphasis on user privacy through 100% on-device processing using models like Qwen, Phi-3, and Gemma.

Main Features

1. Built-in Local AI Chat (/ask): Users initiate a conversational AI session directly within their terminal by typing /ask. The AI operates on-device and is context-aware, meaning it has full visibility of the current terminal session context, including the command history, output, and working directory. It uses local models (Qwen, Phi-3, Gemma) to answer questions about the current state, debug errors, explain output, or plan next steps without any data leaving the Mac.
2. Context-Aware AI Command Suggestions: After executing any command, jebi's AI analyzes the result and the session history to suggest logically subsequent commands. This feature is context-aware, ranking suggestions by relevance to the user's immediate task, aiming to streamline workflow and reduce cognitive load during complex operations.
3. Intelligent Slash Command Palette (/): A built-in toolkit accessible via the / key. This provides instant access to specialized functions without leaving the keyboard-driven flow. Key built-in commands include /ls for an enhanced file explorer with inline previews, /ports for a real-time network port inspector, and /run for a picker for Makefile and npm scripts.
4. Advanced Workspace and UI Management: jebi offers a professional workspace with features like multiple tabs with per-tab accent colors, horizontal or vertical split panes, and prefix-filtered command history using ↑/↓ keys. It includes a beautiful UI with 20+ handcrafted dark color palettes, a grain texture aesthetic, and GPU-accelerated WebGL rendering for smooth performance even with high output volume.
5. Customizable Commands and Productivity Tools: Users can define their own custom slash commands in Preferences, backed by any shell script. This allows for the automation of repetitive workflows. Additional tools include a scrollback search (⌘F), smart prompt segments displaying Git branch or Python environment (clickable to copy), and seamless detection for TUI applications like vim and htop.

Problems Solved

1. Pain Point: Context Switching and Interrupted Flow: Developers frequently lose momentum by having to leave the terminal to search for error explanations or command documentation online. jebi solves this by providing in-terminal error explanation and command suggestion, keeping the user's focus entirely within their coding environment.
2. Pain Point: Privacy Concerns with Cloud AI: Many developer tools require sending code, commands, or context to external APIs for AI processing. jebi addresses this by running all AI inference locally on the user's machine, ensuring that sensitive code, server configurations, and terminal histories never leave their Mac, eliminating the need for API keys or subscriptions.
3. Pain Point: Terminal Workflow Inefficiency: Standard terminals lack intelligent assistance and quick-access utilities. jebi consolidates file browsing, port inspection, and script running into a single, fast, keyboard-driven interface via slash commands, streamlining the development and DevOps workflow.
4. Target Audience: This product is designed for software developers (full-stack, backend, systems), DevOps engineers, SREs, power users, and CLI enthusiasts on macOS who value efficiency, workflow integration, and data privacy. It serves both seasoned professionals who live in the terminal and those who want a more approachable command-line experience.
5. Use Cases: Essential for real-time debugging of failed builds or scripts, where immediate, context-specific explanations are needed. Ideal for managing multiple concurrent projects in split panes, automating daily tasks with custom commands, and securely working with sensitive server or client codebases where cloud AI is prohibited.

Unique Advantages

1. Differentiation: Unlike standard terminals (iTerm2, Terminal.app) or other smart terminals, jebi's primary differentiator is its integrated, session-aware AI that runs entirely locally. It moves beyond basic code completion to provide explanatory intelligence and proactive suggestions, all while guaranteeing complete privacy. Compared to cloud-based AI coding assistants, it eliminates latency, internet dependency, and data security risks.
2. Key Innovation: The core technical innovation is the on-device AI pipeline that can maintain and process full terminal session context in real-time. By leveraging optimized local models (Qwen, Phi-3, Gemma), jebi provides intelligent assistance that understands not just a single command, but the entire narrative of the user's work session, a capability previously reliant on cloud services.

Frequently Asked Questions (FAQ)

1. What AI models does jebi use, and how is it free without a subscription? jebi runs Qwen, Phi-3, and Gemma models locally on your Mac's hardware. It is free because the computation happens on your own device; there are no server costs for the developer, hence no API keys, subscriptions, or cloud fees are required for the end user.
2. Is my terminal history and data private when using jebi's AI features? Yes, complete privacy is guaranteed. All AI processing, including the /ask chat and command suggestions, occurs 100% locally on your machine. Your commands, outputs, and working directory context are never transmitted to any external server.
3. How do I install jebi on my Mac? jebi is installed via Homebrew. Open your terminal and run brew tap jebi-sh/tap followed by brew install --cask jebi. Upon first launch, you may need to go to System Settings → Privacy & Security and click "Open Anyway" to bypass macOS security blocks.
4. What are the hardware requirements for running the local AI models smoothly? While exact specs depend on the models used, jebi is optimized for modern Apple Silicon Macs (M1/M2/M3/M4). The AI features are designed to be responsive without excessive resource consumption, but performance will be best on machines with at least 16GB of unified memory for optimal model loading and inference speed.
5. Can I extend jebi with my own custom commands and workflows? Absolutely. jebi supports user-defined custom slash commands. You can create commands in the Preferences that execute shell scripts or dynamic prompts, allowing you to automate and personalize your workflow while maintaining seamless integration with the built-in tools like /ls and /ports.