1585 Platform Setup A4512e9b

Platform Setup (Airbyte Cloud / Hosted Mode)

This guide covers setting up connectors through the Airbyte Platform (app.airbyte.ai), where connectors appear in your dashboard and credentials are managed securely.

When to Use Platform Mode

Use platform mode when you want:

• Connectors visible in the Airbyte UI
• Managed credential storage and rotation
• Entity cache for faster repeated queries
• Multi-tenant SaaS deployments
• OAuth flows handled server-side

Prerequisites

Get Your Airbyte Platform Credentials

1. Sign up at app.airbyte.ai
2. Go to Settings > API Keys
3. Create a new API key to get:
   • AIRBYTE_CLIENT_ID
   • AIRBYTE_CLIENT_SECRET

These are your application-level credentials for creating and managing connectors programmatically.

Define Your External User ID

The external_user_id is YOUR identifier for the user or tenant - you define it:

• "user_123" - for user-scoped connectors
• "acme-corp" - for organization-scoped connectors
• "tenant_abc" - for multi-tenant applications

This ID is used to:

• Scope connectors to specific users/tenants
• Look up existing connectors without storing connector IDs

Creating Connectors with create_hosted()

API Key Connectors

For connectors using API keys (Stripe, Gong, Jira, etc.):

import os
from airbyte_agent_stripe import StripeConnector
from airbyte_agent_stripe.models import StripeAuthConfig

# Create and register a new connector
connector = await StripeConnector.create_hosted(
    external_user_id="user_123",
    airbyte_client_id=os.environ["AIRBYTE_CLIENT_ID"],
    airbyte_client_secret=os.environ["AIRBYTE_CLIENT_SECRET"],
    auth_config=StripeAuthConfig(api_key=os.environ["STRIPE_API_KEY"]),
    name="Stripe Source"  # Optional: display name in UI
)

# Connector is created and ready to use programmatically
# For UI visibility, register a template (see "Making Connectors Visible in the UI" below)

# Use it immediately
result = await connector.execute("customers", "list", {"limit": 10})

Gong Example

from airbyte_agent_gong import GongConnector
from airbyte_agent_gong.models import GongAccessKeyAuthenticationAuthConfig

connector = await GongConnector.create_hosted(
    external_user_id="acme-corp",
    airbyte_client_id=os.environ["AIRBYTE_CLIENT_ID"],
    airbyte_client_secret=os.environ["AIRBYTE_CLIENT_SECRET"],
    auth_config=GongAccessKeyAuthenticationAuthConfig(
        access_key=os.environ["GONG_ACCESS_KEY"],
        access_key_secret=os.environ["GONG_ACCESS_KEY_SECRET"]
    ),
    name="Gong Source"
)

GitHub with Personal Access Token

from airbyte_agent_github import GithubConnector
from airbyte_agent_github.models import GithubPersonalAccessTokenAuthConfig

connector = await GithubConnector.create_hosted(
    external_user_id="user_123",
    airbyte_client_id=os.environ["AIRBYTE_CLIENT_ID"],
    airbyte_client_secret=os.environ["AIRBYTE_CLIENT_SECRET"],
    auth_config=GithubPersonalAccessTokenAuthConfig(
        token=os.environ["GITHUB_TOKEN"]
    ),
    name="GitHub Source"
)

Using Existing Connectors

Once a connector is created, you don't need to pass auth_config again:

# Option A: Look up by external_user_id
connector = StripeConnector(
    external_user_id="user_123",
    airbyte_client_id=os.environ["AIRBYTE_CLIENT_ID"],
    airbyte_client_secret=os.environ["AIRBYTE_CLIENT_SECRET"],
)

# Option B: Use connector_id directly (if you have it cached)
connector = StripeConnector(
    connector_id="your-connector-uuid",  # From previous create_hosted() or API
    airbyte_client_id=os.environ["AIRBYTE_CLIENT_ID"],
    airbyte_client_secret=os.environ["AIRBYTE_CLIENT_SECRET"],
)

# Operations use credentials stored in Airbyte
result = await connector.execute("customers", "list", {"limit": 10})

OAuth Connectors (Salesforce, HubSpot, Google Drive, etc.)

OAuth connectors require a server-side flow to capture user authorization.

Step 1: Initiate OAuth

from airbyte_agent_salesforce import SalesforceConnector

# Start the OAuth flow
oauth_response = await SalesforceConnector.initiate_oauth(
    external_user_id="user_123",
    airbyte_client_id=os.environ["AIRBYTE_CLIENT_ID"],
    airbyte_client_secret=os.environ["AIRBYTE_CLIENT_SECRET"],
    redirect_url="https://yourapp.com/oauth/callback"
)

# Redirect user to the consent URL
print(f"Redirect user to: {oauth_response.consent_url}")

Step 2: Handle OAuth Callback

When the user completes authorization, they're redirected to your callback URL with a secret_id parameter:

https://yourapp.com/oauth/callback?secret_id=abc123

Step 3: Create Connector with OAuth Secret

connector = await SalesforceConnector.create_hosted(
    external_user_id="user_123",
    airbyte_client_id=os.environ["AIRBYTE_CLIENT_ID"],
    airbyte_client_secret=os.environ["AIRBYTE_CLIENT_SECRET"],
    server_side_oauth_secret_id="abc123",  # From callback
    name="Salesforce Source"
)

OAuth Connectors Reference

Making Connectors Visible in the UI

After create_hosted() succeeds, the connector is functional programmatically. To make it appear in the Airbyte UI's Connectors page, you must register it as a template.

Register a UI Template

curl -X POST 'https://api.airbyte.ai/api/v1/integrations/templates/sources' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer <APPLICATION_TOKEN>' \
  -d '{
    "actor_definition_id": "<CONNECTOR_DEFINITION_ID>",
    "name": "Gong",
    "partial_default_config": {},
    "mode": "DIRECT"
  }'

Parameters

Get the Application Token

curl -X POST 'https://api.airbyte.ai/api/v1/account/applications/token' \
  -H 'Content-Type: application/json' \
  -d '{
    "client_id": "<AIRBYTE_CLIENT_ID>",
    "client_secret": "<AIRBYTE_CLIENT_SECRET>"
  }'

Use the access_token from the response as <APPLICATION_TOKEN>.

Example: Register Gong Connector

curl -X POST 'https://api.airbyte.ai/api/v1/integrations/templates/sources' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer eyJhbGciOiJIUzI1...' \
  -d '{
    "actor_definition_id": "32382e40-3b49-4b99-9c5c-4076501914e7",
    "name": "Gong",
    "partial_default_config": {},
    "mode": "DIRECT"
  }'

After registration, your connector appears in the Connectors page with a card showing the name and "Direct" badge.

Creating Connector Instances

After registering a template, you need to create an actual connector instance to access data.

Important: The SDK's create_hosted() has a known bug with the API URL (uses /v1/integrations/connectors instead of /api/v1/integrations/connectors). Use the HTTP API directly until this is fixed upstream.

Step 1: List Workspaces

First, find your workspace name:

curl 'https://api.airbyte.ai/api/v1/workspaces' \
  -H 'Authorization: Bearer <APPLICATION_TOKEN>'

Use the workspace name (not id) as your external_user_id when creating connectors.

Step 2: Create Instance

curl -X POST 'https://api.airbyte.ai/api/v1/integrations/connectors' \
  -H 'Authorization: Bearer <APPLICATION_TOKEN>' \
  -H 'Content-Type: application/json' \
  -d '{
    "workspace_name": "<YOUR_WORKSPACE_NAME>",
    "connector_definition_id": "<DEFINITION_ID>",
    "name": "my-connector",
    "credentials": {
      // Connector-specific fields ONLY - no auth_type
    }
  }'

Credentials Format by Connector

Do NOT include auth_type - the API infers it from the credentials structure.

Step 3: Verify Connection

curl -X POST 'https://api.airbyte.ai/api/v1/integrations/connectors/<CONNECTOR_ID>/execute' \
  -H 'Authorization: Bearer <APPLICATION_TOKEN>' \
  -H 'Content-Type: application/json' \
  -d '{"entity": "users", "action": "list", "params": {"limit": 1}}'

A successful response confirms your connector is working.

Verifying Connector in UI

After registering the template:

1. Go to app.airbyte.ai
2. Navigate to Connectors page
3. Your connector should appear with the name you specified and a "Direct" badge
4. Click to view status, configuration, and usage

Entity Cache

Platform connectors support entity caching for faster repeated queries. Entity cache is managed through the Airbyte UI:

1. Go to app.airbyte.ai
2. Select your connector
3. Enable entity cache in the connector settings

When entity cache is enabled:

• Common queries are cached and served faster
• Cache syncs automatically in the background
• View cache status in the Airbyte UI

Environment Variables Setup

Create a .env file for your platform credentials:

# Airbyte Platform (from app.airbyte.ai > Settings > API Keys)
AIRBYTE_CLIENT_ID=your_client_id
AIRBYTE_CLIENT_SECRET=your_client_secret

# Connector credentials (varies by connector)
STRIPE_API_KEY=sk_live_...
GONG_ACCESS_KEY=...
GONG_ACCESS_KEY_SECRET=...
GITHUB_TOKEN=ghp_...

Load in Python:

from dotenv import load_dotenv
import os

load_dotenv()

# Now use os.environ to access credentials

Troubleshooting Platform Setup

"Invalid client credentials"

• Verify AIRBYTE_CLIENT_ID and AIRBYTE_CLIENT_SECRET are correct
• Regenerate API keys in app.airbyte.ai > Settings > API Keys

"Connector not appearing in UI"

• Ensure create_hosted() completed without errors
• Check the external_user_id matches your workspace
• Refresh the UI page

"OAuth flow failing"

• Verify redirect URL matches exactly
• Check OAuth app settings in the third-party service
• Ensure required scopes are configured

"create_hosted() not found"

• Update your SDK: pip install --upgrade airbyte-agent-{connector}
• create_hosted() was added in SDK v0.1.0

Related Documentation

• OSS Setup - Local SDK without platform integration
• Authentication - Auth patterns by connector
• Programmatic Setup - HTTP API and curl examples
• Troubleshooting - Common errors and solutions
• Entity-Action API - Core API patterns