cmdbox

cmdbox

cmdbox is a command foundation for AI system development that lets you launch with minimal implementation. A single feature implementation can be exposed through CLI / REST API / MCP SV / Web UI / Edge UI / remote execution.

• Documentation: https://hamacom2004jp.github.io/cmdbox/

Why cmdbox for AI System Developers

• No repeated reimplementation
  • Implement a Feature command once, then use the same capability via CLI, Web UI, and REST API.
  • It is easy to promote a PoC CLI workflow into operational Web/API interfaces without rewriting business logic.
• AI-ready components out of the box
  • Modes such as agent, llm, mcpsv, and a2asv let you manage agent execution, model operations, MCP exposure, and A2A integration on one platform.
  • You can separate tool execution, inference, knowledge retrieval, and external integration while keeping a consistent operational model.
• Rich Web UI for daily operations
  • Beyond APIs and CLI, cmdbox provides operational screens for commandlets, agents, filer, limiter, audit, and user management.
  • Teams can run, monitor, and control workflows from a browser without building a separate admin console.
• Designed for remote execution
  • Server-side commands can run through Redis, which fits node separation and scale-out architectures.
  • Heavy workloads can be offloaded to servers while keeping client components lightweight.
• Fast, configuration-driven extensibility
  • features.yml defines command discovery, aliases, argument rules, and agent execution rules, separating implementation from operations policy.
  • You can adjust exposed commands and execution policies without code changes, which is useful for environment-specific control.
• Built-in governance capabilities
  • You can combine audit logging, authentication/authorization (signin / oauth2 / saml / rule settings), and limiter controls.
  • It is easier to trace who executed which command under which conditions, enabling enterprise-grade operational control.

Problems It Solves Well

• You need to launch internal AI tools quickly
  • Start with CLI features, then expose the same capabilities internally via Web and REST API.
• Multiple teams/environments need one shared foundation
  • Run the same codebase across dev/stg/prod while switching execution policies through configuration.
• You need safe integration between agents and existing systems
  • Combine MCP/A2A integration with authorization rules to control the scope of agent actions.
• You need usage control and cost governance for AI operations
  • Use Limiter settings to enforce quotas/rate rules, prevent abuse, and keep usage and billing behavior predictable.
  • Define reusable plans and billing policies (period-based or metered) to align access control with commercial and internal chargeback models.

What You Can Build

• Custom command development
  • Support complex option definitions, client/server execution, and Web execution.
  • Organize functionality by mode / cmd and standardize execution through a shared option model.
• LLM integration
  • Chat, embedding, and configuration save/load/list workflows.
  • Manage model configurations and treat inference calls as operational commands.
• Agent integration
  • Save/load agent settings, integrate MCP servers, and connect A2A servers.
  • Expand agent capabilities in phases while keeping clear boundaries on exposed tools.
• Web UI operations
  • Use built-in screens for commandlets, agents, filer, limiter, audit, and user management.
  • Manage day-to-day operations from the browser while keeping the same backend command model.
• Operational controls
  • Audit logs, authentication, command execution policy, and REST API controls.
  • Publish safely with permissions segmented by organization or user group and full traceability.
  • Configure plan and billing operations to support subscription-style access windows and metered charging models.

Typical Use Cases

• Internal knowledge search combining document processing + LLM
• Converting routine operations into commands and running them from Web UI with no-code interaction
• Letting agents call internal tools safely within controlled boundaries
• Running operational administration (filer, limiter, audit, and user control) from the built-in Web UI

Installation

uv pip install cmdbox[app]
cmdbox -v

Docker Environment Setup

• Redis

cmdbox -m cmdbox -c redis_install
cmdbox -m cmdbox -c up -C redis

• PostgreSQL (used by some commands)

cmdbox -m cmdbox -c pgsql_install
cmdbox -m cmdbox -c up -C pgsql

• Full cmdbox suite

cmdbox -m cmdbox -c server_install
cmdbox -m cmdbox -c up -C cmdbox

Quick Start

cmdbox -m server -c start
cmdbox -m web -c start --signin_file .cmdbox/user_list.yml
cmdbox -m mcpsv -c start --signin_file .cmdbox/user_list.yml
cmdbox -m a2asv -c start --signin_file .cmdbox/user_list.yml

Tutorial

The tutorial content has moved to the following documentation:

• docs_src/docs/tutorial.rst
• Published tutorial: https://hamacom2004jp.github.io/cmdbox/docs/tutorial.html
• Full documentation: https://hamacom2004jp.github.io/cmdbox/

Support and Sponsorship

If you find cmdbox useful and would like to support its development, consider becoming a sponsor!

👤 Individual Sponsorship

• $5/month - Recognition in README and SPONSORS list

🏢 Corporate Sponsorship

• $100/month (Bronze) - README and SPONSORS list recognition
• $500/month (Silver) - Priority issue resolution
• $1000/month (Gold) - Custom feature consultation

See SPONSORS.md for details.

Your sponsorship helps us continue developing and maintaining cmdbox!

Contributing

We welcome contributions! Please feel free to:

• Report bugs via GitHub Issues
• Submit pull requests with improvements
• Share your use cases and feedback

License

This project is licensed under the MIT License, see the LICENSE file for details