Blog

AI Agents

The Airbyte Agent CLI is here

Airbyte Agent CLI is here: build, run, and manage AI-powered data agents from your terminal with Airbyte Agents.

AI AGENTS

June 1, 2026

Michel Tricot

Summarize with AI:

Agents are moving into the terminal.

Claude Code and Codex CLI have allowed AI systems to work outside of chat interfaces and into where developers work: the terminal.

Agents living in the terminal have to interact with infrastructure that cuts directly into their work. There's no need for an additional dashboard. They just need to know what tools are available and be able to act upon them.

Today, we’re launching the Airbyte Agent CLI, a standalone binary that brings the power of Airbyte Agents to your terminal.

Why did we build a CLI?

Because every framework for agents runs into the same problem. Agents can reason extremely well but are otherwise blocked when they attempt to interact with systems outside of themselves.

Airbyte Agents were built to remove that friction. The CLI presents a simplified interface to the platform without the need for an SDK or Python runtime.

Getting Started

Installation of the CLI is as easy as one command:

curl -fsSL https://airbyte.ai/install.sh | bash

For macOS users, installation via Homebrew is also available:

brew install airbytehq/tap/airbyte-agent-cli

After installation, authenticate with the CLI using:

airbyte-agent login

And list your workspaces using:

airbyte-agent workspaces list

All commands accept JSON input and JSON output, enabling easy integration with existing scripts and automation tools.

Agents that teach themselves your data

Airbyte Agents learn about the tools that they are meant to interact with via the schema command.

airbyte-agent schema connectors execute

The CLI returns the structure of the command that can be executed by the agent. This allows for agents to explore what tools are available and what actions they can take without being trained on documentation for those tools.

Agents can discover a tool, a tool describes itself, and an agent acts. This is a loop that enables the agent to function natively within the CLI.

Four interfaces, one platform

The CLI gives developers four ways to interact with Airbyte Agents. Each interface is meant for specific situations.

The SDK provides deep integrations for Python-native agent frameworks. The Agents API allows frameworks to interact with Airbyte Agents in a custom way. The MCP provides support for those who prefer to build agentic workflows directly inside their LLM client. And finally, the CLI enables terminal workflows and CI/CD pipelines.

They all connect to the same platform with the same connectors and context store for agents to interact with.

The interface best suited to your framework or platform is the one that you need. The infrastructure remains the same.

Get started

If you're wiring up Airbyte Agents to an existing agent framework, the install script will install the Airbyte Agent skill along with the CLI tool. This will give your agent the tools it needs to begin reading from and writing to your connected systems.

The source code and releases for the Airbyte Agent CLI can be found in the Airbyte Agent CLI repository. The full documentation is available here.

The terminal is where agents will live in the future. We brought Airbyte Agents there for you.

Build something with the CLIU, and let us know what you think!

Integrate with 700+ apps using Airbyte

Move data from 700+ sources into warehouses, lakes, and beyond. Set up pipelines in minutes with pre-built connectors and the Connector Builder.

Try it free Talk to sales

About the Author

Michel Tricot

Michel Tricot is the CEO and Co-Founder of Airbyte.

Integrate with 700+ apps using Airbyte

Try Airbyte for free

The Airbyte Agent CLI is here

Integrate with 700+ apps using Airbyte

Integrate with 700+ apps using Airbyte

Related posts