All Topics
All Topics
Technology
Technology
AI
AI
Business
Business
Entertainment
Entertainment
News
News
Programming
Programming
Science
Science
Design
Design
Environment
Environment
Finance
Finance
Crypto
Crypto
Politics
Politics
Sports
Sports
Education
Education
Gaming
Gaming
Art
Art
Music
Music
Health
Health
Security
Security
Books
Books
Food
Food
Travel
Travel
Personal
Personal
Bluesky
Twitter

How Knock built an AI agent for customer messaging using bash and a virtual file system

Knock shares the technical architecture behind their AI agent (Knock Agent) for managing customer messaging resources. The agent was built using bash, a virtual file system (VFS), and their existing management API. The post details how they modeled resources as files in a VFS, used bash as the agent's interface for tool execution, and leveraged the management API for operations. The approach prioritizes simplicity, debuggability, and developer familiarity over complex agent frameworks.

Chris Bell, Knock5h ago11 min readenInsight
Read on knock.app

Key quotes

Our vision was to create an agent that feels natural to use while being deeply integrated with Knock's platform.
By modeling resources as files in a virtual file system, we could leverage the simplicity and power of bash commands.
We chose bash not because it's trendy, but because it's the most universal interface for developers to interact with systems.

From the article

A look at how we architected the Knock Agent with bash, a virtual file system, and the management API.
Continue reading on knock.app

You might also wanna read

Comments

Sign in to join the conversation.

No comments yet. Be the first.