claude-rig

Run multiple Claude Code configurations side by side.

The Problem

Claude Code keeps all configuration in a single ~/.claude/ directory. That works until you need:

• Multiple subscriptions -- You have 2-5 Max accounts and have to log out and back in every time you switch. No way to run two subscriptions at the same time.
• API and subscription side by side -- Want API access for one project and your Max subscription for another? Same deal -- one config, one auth.
• Different tools per project -- Your web project needs Tailwind skills and a database MCP server. Your CLI tool needs none of that. But every session loads the same plugins, MCP servers, and hooks.
• Testing without risk -- Want to try a new MCP server without it leaking into your production setup? You'd have to disable everything else, test, then re-enable. No isolation.

What claude-rig Does

Each configuration becomes its own isolated rig -- separate settings, skills, plugins, agents, commands, hooks, MCP servers, and instructions. Auth can be shared across rigs or kept separate. Run them side by side in different terminals.

# Terminal 1: your Max subscription, minimal setup
claude --rig=minimal

# Terminal 2: different Max account, full web dev stack
claude --rig=webdev

# Terminal 3: API access, experimental MCP server you're testing
claude --rig=experiment

Why claude-rig

Most tools in this space swap credential files between accounts. claude-rig isolates the entire configuration -- and lets you run them at the same time.

	claude-rig	Credential switchers
Run multiple configs simultaneously	Yes	No -- must close Claude first
Isolated settings, plugins, MCP, hooks	Yes	No -- auth only
Choose what's shared vs isolated per rig	Yes	No
Per-project auto-selection	Yes	No
Status, diff, export/import	Yes	No
Dependencies	None (single binary)	Varies

What It Doesn't Do

• Doesn't modify Claude Code. Uses the official CLAUDE_CONFIG_DIR environment variable and --add-dir flag. No patches, no forks.
• Doesn't manage Claude Code installation. Install Claude Code yourself -- claude-rig manages its configuration.
• Doesn't replace project-level config. Your project's CLAUDE.md and .claude/ still work normally. Rigs handle user-level configuration.
• No external dependencies. Single Go binary, stdlib only.

Install

Homebrew (macOS / Linux):

brew install edimuj/tap/claude-rig

Shell script (macOS / Linux):

curl -sSL https://raw.githubusercontent.com/edimuj/claude-rig/main/install.sh | sh

PowerShell (Windows):

irm https://raw.githubusercontent.com/edimuj/claude-rig/main/install.ps1 | iex

Go install:

go install github.com/edimuj/claude-rig/cmd/claude-rig@latest

Download binaries: GitHub Releases -- macOS, Linux, Windows.

Quick Start

# One-time setup
claude-rig init

# Create your first rig
claude-rig create minimal

# Or clone your existing ~/.claude/ config as a starting point
claude-rig clone default webdev --link-auth

# Launch explicitly
claude-rig launch webdev

# Or bind to a project and just run claude-rig
cd ~/projects/my-app
claude-rig rc webdev
claude-rig          # picks up the rig automatically

Commands

Command	Description
blueprint <sub>	Shareable rig specs: create, apply, inspect, list, pack
clone <src|default> <dest>	Clone a rig or ~/.claude/ config
create <name>	Create a new rig (--link-auth, --no-isolate-defaults, --isolate-all)
delete <name>	Delete a rig
diff <rig1> <rig2>	Compare two rigs
doctor [--fix]	Diagnose and fix broken symlinks, plugin/MCP health
export <rig> [file]	Export rig to .tar.gz
import <file> <name>	Import rig from archive
inherit <items> [rig]	Inherit global skills/agents/hooks/commands
init	Initialize claude-rig and install shell integration
isolate <rig> <items>	Isolate items per rig
isolation [rig]	Show isolation status for one or all rigs
launch [name] [args]	Launch Claude Code with a rig
link-auth <name>	Link rig to shared auth
list	List all rigs (* = running sessions)
pin <version>	Pin rig to a specific Claude Code version
plugin <subcommand>	Run claude plugin commands on the correct rig
rc [name]	Show or create .claude-rig file for current directory
rename <old> <new>	Rename a rig
set-args [name] <args>	Set default launch args (global or per-rig)
settings <sub>	Manage default settings applied to all rigs
share <rig> <items>	Reverse isolation
show-args [name]	Show default launch args
status [rig]	Show rig status: disk, sessions, last used
sync [rig]	Sync symlinks, inherited items, plugins, MCP, settings
uninherit <items> [rig]	Stop inheriting
unlink-auth <name>	Remove shared auth
unpin	Remove version pin
update	Update Claude Code
update-plugins [rigs]	Update marketplace plugins across rigs
versions	List installed Claude Code versions on disk

How It Works

Each rig is a full config directory under ~/.claude-rig/rigs/<name>/:

~/.claude-rig/rigs/webdev/
    .claude.json            # real file (MCP servers, onboarding state)
    CLAUDE.md               # real file (rig-specific instructions)
    settings.json           # real file (rig-specific config)
    rig.json                # isolation config, synced plugins/MCP, overrides
    skills/                 # real directory
    plugins/                # real directory (synced from global by default)
    conversations/          # real dir (isolated by default)
    history.jsonl           # real file (isolated by default)
    telemetry/ -> ~/.claude/ # symlink (shared by default)
    ...

On launch, claude-rig resolves the binary (pinned or latest on disk), sets CLAUDE_CONFIG_DIR to the rig directory, loads global ~/.claude/CLAUDE.md via --add-dir, refreshes symlinks, then replaces itself with Claude Code via exec. Two instances with different rigs can run at the same time without conflicts.

Blueprints -- Share Your Setup

You've spent hours dialing in your Claude Code setup. The right plugins, MCP servers, skills, settings, CLAUDE.md instructions. Then someone asks "what's your setup?" and you have nothing to hand them. No export, no shareable format. You end up writing a wall of text listing everything by hand.

Blueprints fix that. One command extracts your rig into a portable spec. Another recreates it from scratch.

# Extract your setup
claude-rig blueprint create my-setup --from go

# Anyone can recreate it -- from GitHub, a file, or a local directory
claude-rig blueprint apply edimuj/my-setup --link-auth

Unlike export/import (which creates full file archives for backup), blueprints are lightweight specs. Plugins are listed as install references, not binary blobs. The whole thing is a small JSON file plus any custom skills or agents you've written. Push it to GitHub and your followers, teammates, or future self can spin up your exact setup in one command.

What you can do with blueprints:

• Share your setup on social media -- "here's the rig I use for all my Go projects"
• Onboard new team members with one command instead of a wiki page
• Reproduce your setup on a new machine without remembering what you installed
• Publish alongside a blog post or tutorial so readers can follow along with your exact config
• Keep a library of blueprints for different types of work -- web, mobile, data, DevOps

See docs/blueprints.md for the full reference.

Platform Support

• Linux -- Full support
• macOS -- Full support
• Windows -- Requires Developer Mode (for symlinks). See docs/windows.md.

Documentation

Topic	Doc
Per-project RC files, shell integration, CLAUDE.md, templates	docs/configuration.md
Isolation model, configurable isolation, global inheritance	docs/isolation.md
Plugin sync, MCP sync, default settings, version management	docs/sync.md
Blueprints -- share your Claude Code setup with anyone	docs/blueprints.md
Export, import, status, diff	docs/export-import.md
Windows setup, known limitations	docs/windows.md

Also Check Out

Project	Description
tokenlean	Lean CLI tools for AI agents -- reduce context, save tokens
claude-mneme	Automatic session memory -- every session picks up where the last left off
vexscan	Security scanner for AI agent plugins, skills, MCPs, and configs
claude-workshop	Collection of useful plugins and tools for Claude Code
claude-simple-status	No-frills statusline showing model, context, and quota usage

Development

make build                # build binary
make build-windows        # cross-compile for Windows (amd64)
make run ARGS="version"   # run without installing
make install              # install to ~/go/bin/

License

MIT