Quick Start

This guide gets ByteMind installed and running your first AI coding task in about 5 minutes.

Prerequisites

ByteMind ships as a pre-compiled binary — no Go installation required.

Requirement	Details
OS	macOS, Linux, or Windows
API Key	Any OpenAI-compatible service or Anthropic
Network	Access to your LLM provider endpoint

Step 1: Install

macOS / Linux

curl -fsSL https://raw.githubusercontent.com/1024XEngineer/bytemind/main/scripts/install.sh | bash

Windows (PowerShell)

iwr -useb https://raw.githubusercontent.com/1024XEngineer/bytemind/main/scripts/install.ps1 | iex

Verify the installation:

bytemind --version

Install Location

Defaults to ~/bin (Linux/macOS) or %USERPROFILE%\bin (Windows). If the command is not found, make sure that directory is on your PATH.

Step 2: Create a Config

Create a .bytemind/ directory inside your project:

mkdir -p .bytemind

Create .bytemind/config.json with your provider settings:

{
  "provider": {
    "type": "openai-compatible",
    "base_url": "https://api.openai.com/v1",
    "model": "gpt-4o",
    "api_key": "YOUR_API_KEY"
  },
  "approval_policy": "on-request",
  "max_iterations": 32,
  "stream": true
}

Keep secrets out of Git

Never commit a config file with a real api_key. Add .bytemind/ to your .gitignore, or use the api_key_env field to read the key from an environment variable instead.

Key fields at a glance:

Field	Description	Default
provider.type	openai-compatible or anthropic	openai-compatible
provider.base_url	API endpoint URL	https://api.openai.com/v1
provider.model	Model ID to use	gpt-5.4-mini
provider.api_key	API key (plain text)	—
provider.api_key_env	Env var name to read the key from	BYTEMIND_API_KEY
approval_policy	When to prompt for approval	on-request
max_iterations	Max tool-call rounds per task	32
stream	Enable streaming output	true

See Config Reference for the full field list.

Step 3: Start Chat Mode

From your project directory, run:

bytemind chat

ByteMind reads .bytemind/config.json from the current directory, initializes a session, and enters interactive mode.

Sessions are auto-saved

Every conversation is persisted. Next time you run bytemind chat, use /sessions to list previous sessions and /resume <id> to continue one.

Step 4: Run Your First Task

Try one of these starter prompts:

Fix failing tests

Find all failing unit tests, analyze the root cause, and fix them with minimal changes.

Understand the codebase

Walk me through this project's directory structure and main entry points. Produce a summary.

Investigate a bug with a skill

/bug-investigation symptom="login endpoint returns 500"

Using Skills

Slash commands starting with / activate built-in skills that inject domain-specific instructions into the agent. For example, /bug-investigation guides the agent through a structured diagnosis workflow. Type /help to see all available commands.

Session Commands

Command	Description
/help	Show all available commands
/session	Show current session details
/sessions	List recent sessions
/resume <id>	Resume a session by ID or prefix
/new	Start a new session
/quit	Exit

Next Steps

• Installation — version pinning, build from source
• Configuration — Anthropic, custom endpoints, sandbox
• Core Concepts — modes, sessions, approval policy
• Chat Mode — best practices and workflows