OpenAI

OpenAI Platform is a developer platform for building applications using OpenAI AI models and APIs.

The OpenAI connector syncs data from OpenAI Platform and does not include data from ChatGPT.

Features

Feature Name	Supported	Notes
Capture deletes		`USERS`, `PROJECT`, and `PROJECT_USER` tables.
History mode
Custom data
Data blocking		Column level
Column hashing
Re-sync		Connection level and table level. A parent table re-sync includes all associated child tables.
Row filtering
API configurable		API configuration
Priority-first sync
Fivetran data models
Private networking
Authorization via API

Supported deployment models

We support the SaaS and Hybrid deployment models for the connector.

You must have an Enterprise or Business Critical plan to use the Hybrid Deployment model.

Setup guide

Follow our step-by-step OpenAI setup guide to connect OpenAI Platform with your destination using Fivetran connectors.

Sync overview

To capture new records and updates, we incrementally sync the following tables:

AUDIT_LOG
AUDIO_SPEECH
AUDIO_TRANSCRIPTION
CODEX_CODE_REVIEW
CODEX_CODE_REVIEW_RESPONSE
CODEX_USAGE
COMPLETION
COST
EMBEDDING

We re-import the remaining tables during every sync.

Schema information

This schema applies to all OpenAI connections.

Explore OpenAI schema ERD

Schema notes

We sync data from the ChatKit API and Vector Stores API into the following tables:

CHATKIT_THREAD
CHATKIT_THREAD_ITEM
CHATKIT_THREAD_ITEM_ATTACHMENT
CHATKIT_THREAD_ITEM_CONTENT
CHATKIT_THREAD_ITEM_TASK
VECTOR_STORE
VECTOR_STORE_FILE

These APIs are currently in beta and may undergo changes. If OpenAI modifies or discontinues the APIs in the future, we will update the schema to reflect those changes.

API key data scopes

Fivetran syncs data from OpenAI using the API keys you provide during connection setup. Each API key provides access to a different scope of data and can be used independently. You can configure only the API keys for the data that you want to sync. For example, if you only want to sync data from the Administration API, you only need to provide an Admin API key.

API key	Data scope	Data synced
Admin API key	OpenAI organization-level data	Organization administration data, including users, roles, groups, permissions, projects, service accounts, audit logs, Admin API keys, certificates, and invites. Fivetran also uses this key to sync organization-wide usage and cost reporting data, including completions, embeddings, audio usage, and cost metrics.
Project-specific API key	OpenAI Platform project-level data	Project-scoped data from the selected OpenAI Platform project, including files, vector stores, containers, fine-tuning jobs, batches, evaluations, chat completions, ChatKit threads, skills, video resources, and audio voice consent records.
Codex Enterprise Analytics API key	ChatGPT workspace-level Codex analytics data	Workspace-level and per-user Codex analytics, including usage, credits, code review activity, review comment priorities, reactions, replies, and engagement metrics. This key does not provide access to OpenAI Platform organization-level administration, usage, or cost data.