Quick Navigation: Migration Guide • New Direction • Resources

This repository has been restructured. Core packages have migrated to @airsprotocols.

• Status: Published on crates.io v1.0.0-rc.1
• Repository: https://github.com/airsstack/airsprotocols/tree/main/protocols/mcp
• Documentation: https://airsstack.github.io/airsprotocols/protocols/mcp/
• Install: cargo add airsprotocols-mcp

• Status: Published on crates.io v1.0.0-rc.1
• Repository: https://github.com/airsstack/airsprotocols/tree/main/mcp/servers/filesystem
• Documentation: https://airsstack.github.io/airsprotocols/servers/filesystem/
• Install: cargo install airsprotocols-mcpserver-filesystem

• Status: Project discontinued
• Reason: No longer maintained
• Migration Path: None available

AirsStack is evolving into a unified CLI controller for AI infrastructure management, inspired by kubectl's approach to Kubernetes orchestration.

A single, powerful command-line tool that serves as the control plane for:

• AI Stack Management: Deploy, configure, and manage complete AI stacks
• Component Orchestration: Lifecycle management for WASM components and services
• Protocol Integration: Seamless interaction with MCP, A2A, and emerging AI protocols
• Build Tooling: Unified build and deployment system for the airsstack ecosystem

WASM-First Design

AirsStack is built on a WebAssembly-first architecture, leveraging @airssys/airssys-wasm as the foundation:

• Agents as WASM Components: All AI agents are packaged as WASM binaries, ensuring portability, security, and performance
• Plugin Architecture: Components and plugins are WASM modules that can be loaded dynamically into the AirsStack runtime
• Host-Component Model: The airsstack CLI acts as the WASM host, managing multiple components/plugins with isolation and security
• Inter-Component Communication: WASM components can communicate with each other through the runtime's message-passing system

Specification Framework

Built on @airsdlc (AI Development Lifecycle) as the core framework:

• Standardized Specifications: All configurations follow AI-DLC artifact guidelines (PRD, RFC, ADR, DAA, Playbooks, TIPs)
• Lifecycle Management: Comprehensive development lifecycle support from planning to deployment
• Extensible Artifacts: Support for custom artifact types and workflows
• Attribution System: Built-in attribution and provenance tracking

Inspired by kubectl

• Single entry point for all operations
• Declarative configuration management
• Extensible plugin architecture through WASM
• Consistent developer experience

WASM-Native

• All agents run as isolated WASM components
• Sandboxed execution with fine-grained permissions
• Cross-platform compatibility (Linux, macOS, Windows, cloud)
• Hot-reload and zero-downtime updates

Practical & Visionary

• Start with immediate developer needs
• Build toward comprehensive AI infrastructure orchestration
• WASM-first for portability and performance
• Community-driven extensibility through plugin ecosystem

airsstack [command] [resource] [flags]

Commands:
  stack       Manage AI stacks and configurations
  component   Orchestrate WASM components and services
  server      Control MCP and protocol servers
  build       Build and package airsstack projects

# Manage MCP servers
airsstack server start filesystem --config ./config.toml
airsstack server list
airsstack server logs filesystem

# Deploy AI stacks
airsstack stack apply -f my-stack.yaml
airsstack stack get my-stack

# Manage WASM components
airsstack component deploy my-agent.wasm
airsstack component list

Update your Cargo.toml:

# Old
[dependencies]
airs-mcp = "0.1.0"

# New
[dependencies]
airsprotocols-mcp = "1.0.0-rc.1"

Update your imports:

// Old
use airs_mcp::*;

// New
use airsprotocols_mcp::*;

Install the new version:

# Old
cargo install airs-mcpserver-fs

# New
cargo install airsprotocols-mcpserver-filesystem

Update Claude Desktop config:

{
  "mcpServers": {
    "filesystem": {
      "command": "airsprotocols-mcpserver-filesystem",
      "args": ["serve", "--allowed-path", "/path/to/project"]
    }
  }
}

• This project has been discontinued
• No direct replacement available
• Consider alternative memory/context management solutions

AirsStack Controller is part of a larger ecosystem of interconnected projects:

graph TB
    subgraph Foundation["Foundation Layer"]
        AIRSDLC["@airsdlc<br/>AI-DLC Framework"]
        AIRSSYS["@airssys/airssys-wasm<br/>WASM Runtime"]
    end
    
    subgraph Controller["AirsStack Controller"]
        AIRSSTACK[AirsStack CLI]
    end
    
    subgraph Protocol["Protocol Layer"]
        AIRSPROTOCOLS["@airsprotocols<br/>Protocols & Servers"]
    end
    
    subgraph Application["Application Layer"]
        AGENTS[AI Agent Components]
        SERVERS[Protocol Servers]
        PLUGINS[Plugin Components]
    end
    
    AIRSDLC -.->|Spec Framework| AIRSSTACK
    AIRSSYS -.->|WASM Runtime| AIRSSTACK
    
    AIRSSTACK --> AIRSPROTOCOLS
    AIRSPROTOCOLS -.->|MCP/A2A| SERVERS
    
    AIRSSTACK -->|Deploy| AGENTS
    AIRSSTACK -->|Manage| SERVERS
    AIRSSTACK -->|Load| PLUGINS
    
    AGENTS -->|Run on| AIRSSYS
    PLUGINS -->|Run on| AIRSSYS
    
    style AIRSSTACK fill:#e3f2fd
    style AIRSDLC fill:#fff3e0
    style AIRSSYS fill:#fff3e0
    style AIRSPROTOCOLS fill:#e8f5e9

@airsdlc - AI Development Lifecycle Framework

• Provides standardized specification and artifact framework
• Defines PRD, RFC, ADR, DAA, Playbooks, and TIPs formats
• Used by AirsStack for spec management and configuration
• Ensures consistent lifecycle management across all tools

@airssys/airssys-wasm - WASM Component Runtime

• Core WASM execution engine for all components
• Provides component isolation, security, and permission management
• Hosts all AI agents as WASM binaries
• Enables plugin architecture through WASM component model
• Powers inter-component communication and message-passing

@airsprotocols - Protocol Implementations

• Production-ready protocol implementations (MCP, A2A)
• Protocol servers managed by AirsStack Controller
• Integration layer between agents and external systems

AirsStack Controller - Orchestration & Management CLI

• Main control plane for the entire ecosystem
• Manages WASM components, agents, servers, and deployments
• Acts as WASM host for plugin and component execution
• Orchestrates multi-agent systems and cloud deployments

• Protocol Documentation: https://airsstack.github.io/airsprotocols/
• MCP Protocol Guide: https://airsstack.github.io/airsprotocols/protocols/mcp/
• Filesystem Server Guide: https://airsstack.github.io/airsprotocols/servers/filesystem/

• airsprotocols (protocols & servers): https://github.com/airsstack/airsprotocols
• airsstack (this repo): https://github.com/airsstack/airsstack

• airsprotocols-mcp: https://crates.io/crates/airsprotocols-mcp
• airsprotocols-mcpserver-filesystem: https://crates.io/crates/airsprotocols-mcpserver-filesystem

The AirsStack Controller is in early conceptual design. We're defining:

• Core command structure and user experience
• Plugin architecture and extensibility model
• Integration patterns with existing protocols
• WASM component management approach

⚠️ Important Note: All development plans outlined below are preliminary and subject to change based on community feedback, technical discoveries, and evolving requirements. This roadmap represents our current vision but should not be considered final.

The following diagram illustrates the planned architecture of AirsStack Controller:

graph TB
    CLI[AirsStack CLI Controller]
    
    subgraph "Core Management Layers"
        SPEC[Spec Management]
        WASM[WASM Component Management]
        SERVER[Server Management]
        AGENT[Agent Management & Orchestration]
        BUILD[Build Management]
        CLOUD[Cloud Management]
    end
    
    subgraph "Infrastructure"
        MCP[MCP Servers]
        PROTO[Protocol Servers]
        COMP[WASM Components]
        AGENTS[AI Agents]
    end
    
    subgraph "Deployment Targets"
        LOCAL[Local Development]
        REMOTE[Remote Servers]
        K8S[Kubernetes]
        CLOUDPROV[Cloud Providers]
    end
    
    CLI --> SPEC
    CLI --> WASM
    CLI --> SERVER
    CLI --> AGENT
    CLI --> BUILD
    CLI --> CLOUD
    
    SPEC -.-> BUILD
    SERVER --> MCP
    SERVER --> PROTO
    WASM --> COMP
    AGENT --> AGENTS
    BUILD -.-> WASM
    BUILD -.-> SERVER
    
    CLOUD --> LOCAL
    CLOUD --> REMOTE
    CLOUD --> K8S
    CLOUD --> CLOUDPROV
    
    AGENT -.-> SERVER
    AGENT -.-> WASM

Configuration and specification management for all AirsStack resources, built on @airsdlc (AI Development Lifecycle) framework.

graph LR
    subgraph "Spec Management"
        VALIDATE[Validate Specs]
        GENERATE[Generate Configs]
        TEMPLATE[Template Engine]
        SCHEMA[Schema Validation]
    end
    
    subgraph "AI-DLC Artifacts"
        PRD[PRD - Requirements]
        RFC[RFC - Proposals]
        ADR[ADR - Decisions]
        DAA[DAA - Analysis]
        PLAY[Playbooks]
        TIP[TIPs - Improvements]
    end
    
    SPECS[Stack Specs] --> VALIDATE
    VALIDATE --> SCHEMA
    TEMPLATE --> GENERATE
    GENERATE --> OUTPUT[Output Configs]
    
    PRD -.-> SPECS
    RFC -.-> SPECS
    ADR -.-> SPECS
    
    style VALIDATE fill:#e1f5ff
    style GENERATE fill:#e1f5ff

Capabilities:

• AI-DLC Framework Integration: All specifications follow standardized AI Development Lifecycle artifacts
• Artifact Types: Support for PRD, RFC, ADR, DAA, Playbooks, and TIPs
• Declarative Configuration: YAML/TOML-based stack definitions
• Spec Validation: Linting and validation against AI-DLC schemas
• Template System: Reusable configuration templates
• Version Management: Schema versioning and migration support
• Attribution Tracking: Built-in provenance and attribution system

Lifecycle management for WebAssembly components, leveraging @airssys/airssys-wasm runtime.

graph TB
    subgraph "WASM Management"
        REGISTRY[Component Registry]
        DEPLOY[Deploy Components]
        LIFECYCLE[Lifecycle Control]
        MONITOR[Monitoring]
    end
    
    subgraph "AirsSys WASM Runtime"
        HOST[WASM Host]
        ISOLATE[Component Isolation]
        BRIDGE[Message Bridge]
        SECURITY[Security & Permissions]
    end
    
    WASM_FILE[.wasm Files] --> REGISTRY
    REGISTRY --> DEPLOY
    DEPLOY --> HOST
    HOST --> ISOLATE
    ISOLATE --> RUNNING[Running Components]
    RUNNING --> LIFECYCLE
    LIFECYCLE --> MONITOR
    MONITOR -.-> LOGS[Logs & Metrics]
    
    BRIDGE -.-> RUNNING
    SECURITY -.-> ISOLATE
    
    style DEPLOY fill:#fff4e1
    style LIFECYCLE fill:#fff4e1
    style HOST fill:#fff4e1

Capabilities:

• WASM Runtime: Built on airssys-wasm for high-performance component execution
• Plugin Architecture: Load components dynamically as plugins into the host runtime
• Component Isolation: Sandboxed execution with capability-based security
• Host-Component Model: AirsStack CLI acts as the WASM host managing multiple components
• Inter-Component Communication: Message-passing between isolated components
• Hot-Reload: Zero-downtime component updates
• Resource Limits: CPU, memory, and I/O quotas per component
• Permission System: Fine-grained capability permissions (network, filesystem, etc.)
• Component Registry: Version management and dependency resolution

Control plane for protocol servers (MCP, A2A, etc.).

graph TB
    subgraph "Server Management"
        DISCOVER[Service Discovery]
        CONTROL[Lifecycle Control]
        CONFIG[Configuration]
        HEALTH[Health Checks]
    end
    
    SERVERS[Protocol Servers] --> DISCOVER
    DISCOVER --> CONTROL
    CONTROL --> START[Start/Stop/Restart]
    CONFIG --> SERVERS
    HEALTH -.-> SERVERS
    
    style CONTROL fill:#e8f5e9
    style HEALTH fill:#e8f5e9

Capabilities:

• Start, stop, restart servers
• Configuration management
• Health monitoring and auto-restart
• Log aggregation
• Port management and routing

Orchestrate AI agents as WASM components from @airssys/airssys-wasm, enabling multi-agent systems with secure inter-agent communication.

graph TB
    subgraph "Agent Orchestration"
        DEPLOY_A[Deploy Agents]
        COORD[Coordination Layer]
        WORKFLOW[Workflow Engine]
        MONITOR_A[Monitoring]
    end
    
    subgraph "WASM Agent Runtime"
        AGENT1[Agent WASM 1]
        AGENT2[Agent WASM 2]
        AGENT3[Agent WASM 3]
        MSG_BUS[Message Bus]
    end
    
    subgraph "AirsStack Controller"
        CLI_CTRL[CLI Controller]
        ORCHESTRATOR[Orchestrator]
    end
    
    AGENT_DEF[Agent Definitions] --> DEPLOY_A
    DEPLOY_A --> AGENT1
    DEPLOY_A --> AGENT2
    DEPLOY_A --> AGENT3
    
    AGENT1 <--> MSG_BUS
    AGENT2 <--> MSG_BUS
    AGENT3 <--> MSG_BUS
    
    MSG_BUS --> COORD
    COORD <--> WORKFLOW
    WORKFLOW --> TASKS[Task Execution]
    MONITOR_A -.-> MSG_BUS
    
    CLI_CTRL --> ORCHESTRATOR
    ORCHESTRATOR -.-> COORD
    ORCHESTRATOR -.-> WORKFLOW
    
    style COORD fill:#f3e5f5
    style WORKFLOW fill:#f3e5f5
    style CLI_CTRL fill:#f3e5f5

Capabilities:

• WASM-Based Agents: All agents packaged as WASM binaries from airssys-wasm
• Multi-Agent Systems: Deploy and coordinate multiple agent instances
• Inter-Agent Communication: Secure message-passing between agent components
• Orchestration Control: AirsStack CLI as the main controller for agent orchestration
• Workflow Engine: Define and execute multi-agent workflows
• Task Distribution: Intelligent task scheduling across agent pool
• State Management: Distributed state coordination
• Failure Recovery: Supervisor trees and restart strategies
• Agent Isolation: Each agent runs in isolated WASM sandbox
• Cloud Deployment: Orchestrate agents across local and cloud environments
• Scalability: Dynamic agent scaling based on workload

Unified build system for AirsStack projects.

graph LR
    subgraph "Build Management"
        SOURCE[Source Code] --> BUILD_SYS[Build System]
        BUILD_SYS --> COMPILE[Compile]
        COMPILE --> TEST[Test]
        TEST --> PACKAGE[Package]
        PACKAGE --> ARTIFACTS[Artifacts]
    end
    
    BUILD_SYS -.-> CACHE[Build Cache]
    ARTIFACTS --> WASM_OUT[WASM Components]
    ARTIFACTS --> BINARY[Binaries]
    
    style BUILD_SYS fill:#fff9c4
    style COMPILE fill:#fff9c4

Capabilities:

• Multi-language support (Rust, Python, JavaScript)
• Incremental builds and caching
• Cross-compilation for WASM
• Dependency management
• Testing integration
• Artifact publishing

Deploy and manage AirsStack resources across environments.

graph TB
    subgraph "Cloud Management"
        PROVISION[Provisioning]
        DEPLOY_C[Deployment]
        SCALE[Scaling]
        MONITOR_C[Monitoring]
    end
    
    subgraph "Targets"
        LOCAL_T[Local]
        DOCKER[Docker]
        K8S_T[Kubernetes]
        AWS[AWS]
        GCP[GCP]
        AZURE[Azure]
    end
    
    PROVISION --> DEPLOY_C
    DEPLOY_C --> LOCAL_T
    DEPLOY_C --> DOCKER
    DEPLOY_C --> K8S_T
    DEPLOY_C --> AWS
    DEPLOY_C --> GCP
    DEPLOY_C --> AZURE
    
    SCALE -.-> DEPLOY_C
    MONITOR_C -.-> DEPLOY_C
    
    style PROVISION fill:#e0f2f1
    style DEPLOY_C fill:#e0f2f1

Capabilities:

• Multi-cloud deployment
• Container orchestration (Docker, Kubernetes)
• Infrastructure as Code
• Auto-scaling policies
• Cost optimization
• Disaster recovery

sequenceDiagram
    participant Dev as Developer
    participant CLI as AirsStack CLI
    participant Spec as Spec Manager
    participant Build as Build System
    participant Deploy as Deployment
    participant Cloud as Cloud Provider
    
    Dev->>CLI: airsstack init my-project
    CLI->>Spec: Generate project specs
    Spec-->>Dev: Project scaffolded
    
    Dev->>CLI: airsstack build
    CLI->>Build: Compile & package
    Build-->>CLI: Artifacts ready
    
    Dev->>CLI: airsstack deploy --env staging
    CLI->>Deploy: Process deployment
    Deploy->>Cloud: Provision resources
    Cloud-->>Deploy: Resources ready
    Deploy->>Cloud: Deploy components
    Cloud-->>CLI: Deployment complete
    CLI-->>Dev: Stack deployed successfully

Phase 1: Foundation

• CLI framework and core architecture
• Basic command structure
• Configuration system
• Spec management foundation

Phase 2: Server Management

• MCP server lifecycle management
• Protocol server orchestration
• Service discovery
• Health monitoring

Phase 3: WASM & Component Management

• WASM component deployment
• Component lifecycle management
• Component registry
• Inter-component communication

Phase 4: Agent Orchestration

• Multi-agent deployment
• Agent coordination layer
• Workflow engine
• State management

Phase 5: Build & Cloud Integration

• Unified build system
• Multi-language support
• Cloud provider integrations
• Kubernetes operator

Phase 6: Ecosystem & Advanced Features

• Plugin system
• Community extensions
• Advanced monitoring and observability
• Cost optimization tools

Current State: Repository restructured, new direction announced
Active Development: Protocol implementations at @airsprotocols
Future Development: AirsStack Controller CLI (planning phase)

✅ MCP Protocol Implementation - Production-ready at @airsprotocols
✅ Filesystem Server - Secure AI-filesystem operations
✅ Comprehensive Documentation - Full guides and API references
🚧 Controller CLI - Coming soon (in planning phase)

While this repository transitions to its new focus, the airsstack ecosystem continues to grow:

• Discussions: https://github.com/orgs/airsstack/discussions
• Issues: Use @airsprotocols for protocol-related issues
• Contributing: Contribution guidelines will be updated as the controller project takes shape

Your AI. Your Data. Your Control.

AirsStack remains committed to:

• Privacy-first AI infrastructure
• Open source and community-driven development
• Composable, interoperable building blocks
• Developer-friendly tooling

Licensed under either of:

• Apache License, Version 2.0 (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
• MIT License (LICENSE-MIT or http://opensource.org/licenses/MIT)

at your option.

Previous State: The pre-restructure codebase is preserved in the archive/pre-restructure-2024-12 branch for reference.

Built with 🦀 Rust | Inspired by kubectl | Powered by the AI Agent Ecosystem