Skip to content
Monk Docs

Architecture

Monk’s architecture consists of several components that work together to provide development and deployment capabilities.

Overview

Local Components

  • VSCode Extension: Provides a chat interface within VSCode and tools that interact with your workspace files using VSCode APIs.

  • Conversational Agent: Runs inside VSCode and interfaces with both local tools and server-side analysis.

  • Monk CLI and Daemon: Non-AI components that handle:

    • Running projects locally
    • Composing Monk packages
    • Cluster management
    • Deployments
    • Runtime operations (logs, status, restarting, scaling)
    • Integration with external deployment platforms

  • Podman: Used for container management on the local system.

Remote Components

  • Multi-Agent Analysis System: Runs on monk.io servers as stateless processes that perform build analysis. This system calls back to local tools to read files, write configurations, and execute approved commands.

Data Flow

  1. The VSCode extension provides the user interface and integration with the editor.
  2. The conversational agent invokes build analysis on monk.io servers.
  3. The analysis system calls tools on the local machine to access files and execute commands.
  4. Container building, secret storage, and runtime operations occur locally.

This architecture ensures that sensitive operations happen locally while leveraging platform-side resources for analysis tasks. Your containers are built locally, your secrets are stored locally, and all operations run through the local CLI and daemon.

CLI Capabilities

The Monk CLI manages:

  • Templates and workloads
  • Container execution
  • Clusters and cloud resources
  • Secrets and registry access

See the CLI Reference for more details.

These components work together to provide a system where most operations run locally while analysis happens remotely in a secure manner.