内容摘录
<div align="center">
<a href="https://airbyte.ai/">
<picture>
<source media="(prefers-color-scheme: dark)" srcset=".github/images/logo-dark.svg">
<source media="(prefers-color-scheme: light)" srcset=".github/images/logo-light.svg">
<img alt="Airbyte Logo" src=".github/images/logo-dark.svg" width="80%">
</picture>
</a>
</div>
<p/>
Python
Publish Connectors
License
Airbyte Stars
Slack
Twitter Follow
Airbyte Agent Connectors are packages that let AI agents call third‑party APIs through strongly typed, well‑documented tools. Each connector is a standalone Python package that you can use directly in your app, plug into an agent framework, or expose through MCP and APIs (coming soon).
What this repo contains
Python SDKs for specific SaaS APIs (for example Gong, Stripe, GitHub).
A consistent connector layout so you can understand and extend any connector quickly.
Ready‑to-use clients that validate auth, handle schemas, and expose typed methods for each operation.
Connector Structure
Each connector is a standalone Python package:
Inside each connector folder, you’ll find:
Generated Python client
Connector-specific README with supported operations
Typed methods generated from Airbyte’s connector definitions
Validation + error handling
When to use these connectors
Use Airbyte Agent Connectors when you want:
Agent‑friendly data access: Let AI agents call real SaaS APIs (e.g., CRM, billing, analytics) with guardrails and typed responses.
Consistent auth and schemas: Reuse a uniform configuration and error‑handling pattern across many APIs. Use connectors inside frameworks like PydanticAI, LangChain, or any custom agent loop.
Composable building blocks: Combine multiple connectors in a single agent to orchestrate multi‑system workflows.
Compared to building ad‑hoc API wrappers, these connectors give you a shared structure, generated clients, and alignment with the rest of the Airbyte ecosystem.
Roadmap
We are actively expanding AI Connectors. Coming soon:
Writes!
MCP support (Model Context Protocol)
Hosted execution, authentication, and search
Demo apps + starter templates
More connectors added weekly
Contributing / Feedback
We actively welcome feedback, ideas, and bug reports.
If you're building AI agents with Airbyte Agent Connectors and want to share ideas or get help, join our community on slack. We'd love to hear what you're building and what connectors or capabilities you'd like to see next.
Claude Code Skills
This repo includes Claude Code skills:
airbyte-agent-connectors: setup and usage guidance for Airbyte Agent Connectors
md2gdoc: convert markdown files to styled Google Docs with template fitting
**Install via plugin marketplace:**
In Claude Code, run:
Then install the plugin:
**Or install manually:**
Once installed, invoke either /airbyte-agent-connectors or /md2gdoc in Claude Code.
See airbyte-agent-connectors and md2gdoc for full usage details.
Tutorial using the Python SDK
In this tutorial, you'll create a new Python project with uv, add a Pydantic AI agent, equip it to use one of Airbyte's agent connectors, and use natural language to explore your data. This tutorial uses GitHub, but if you don't have a GitHub account, you can use one of Airbyte's other agent connectors and perform different operations.
Overview
This tutorial is for AI engineers and other technical users who work with data and AI tools. You can complete it in about 15 minutes.
The tutorial assumes you have basic knowledge of the following tools, but most software engineers shouldn't struggle with anything that follows.
Python and package management with uv
Pydantic AI
GitHub, or a different third-party service you want to connect to
Before you start
Before you begin this tutorial, ensure you have the following.
Python version 3.13 or later
uv
A GitHub personal access token. For this tutorial, a classic token with repo scope is sufficient.
An OpenAI API key. This tutorial uses OpenAI, but Pydantic AI supports other LLM providers if you prefer.
Part 1: Create a new Python project
In this tutorial you initialize a basic Python project to work in. However, if you have an existing project you want to work with, feel free to use that instead.
Create a new project using uv:
This creates a project with the following structure:
Create an agent.py file for your agent definition:
You create .env and uv.lock files in later steps, so don't worry about them yet.
Part 2: Install dependencies
Install the GitHub connector and Pydantic AI. This tutorial uses OpenAI as the LLM provider, but Pydantic AI supports many other providers.
This command installs:
airbyte-agent-github: The Airbyte agent connector for GitHub, which provides type-safe access to GitHub's API.
pydantic-ai: The AI agent framework, which includes support for multiple LLM providers including OpenAI, Anthropic, and Google.
The GitHub connector also includes python-dotenv, which you can use to load environment variables from a .env file.
Part 3: Import Pydantic AI and the GitHub agent connector
Add the following imports to agent.py:
These imports provide:
os: Access environment variables for your GitHub token and LLM API key.
load_dotenv: Load environment variables from your .env file.
Agent: The Pydantic AI agent class that orchestrates LLM interactions and tool calls.
GithubConnector: The Airbyte agent connector that provides type-safe access to GitHub's API.
GithubAuthConfig: The authentication configuration for the GitHub connector.
Part 4: Add a .env file with your secrets
Create a .env file in your project root and add your secrets to it. Replace the placeholder values with your actual credentials.
Add the following line to agent.py after your imports to load the environment variables:
This makes your secrets available via os.environ. Pydantic AI automatically reads OPENAI_API_KEY from the environment, and you'll use os.environ["GITHUB_ACCESS_TOKEN"] to configure the connector in the next section.
Part 5: Configure you…