---
title: OpenAI
description: Connect the OpenAI Administration APIs, Codex, and ChatGPT Enterprise to Coralogix for unified observability across OpenAI.
url: https://coralogix.com/docs/integrations/ai-observability/openai/
product: ai
feature: AI App Catalog
feature_id: ai-app-catalog
feature_identifier: ai/ai-app-catalog
last_updated: '2026-11-04'
---

# OpenAI integrations with Coralogix

Coralogix integrates with OpenAI at the organization level and at the agent level. Pick the surface you want to instrument:

* **[API Platform](https://coralogix.com/docs/integrations/ai-observability/openai/api-platform.md)**: organization-level usage, cost, directory, governance, and audit telemetry pulled from the OpenAI Administration APIs. Requires an OpenAI API Platform organization and an Admin API key.
* **[Codex](https://coralogix.com/docs/integrations/ai-observability/openai/codex.md)**: Codex usage, credits, cost, tool activity, and code review across your ChatGPT Enterprise workspaces, plus a repository-tracking hook for Codex CLI and the ChatGPT desktop app. Requires ChatGPT Enterprise.
* **[ChatGPT](https://coralogix.com/docs/integrations/ai-observability/openai/chatgpt.md)**: ChatGPT Enterprise usage, credits and cost, models, shared GPTs and projects, connectors, and seats per workspace. Requires ChatGPT Enterprise.

All three sources are collected by one **OpenAI Usage & Cost** integration, where each credential independently unlocks its own set of data: an Admin API key for API Platform, a Codex analytics API key per workspace for Codex usage, and a Compliance API key per workspace for ChatGPT and Codex compliance data. Codex and ChatGPT share the Compliance API credentials, so one set of keys serves both. The repository-tracking hook is the only part that runs on developer machines; everything else is pulled from OpenAI's APIs.

Once collected, Codex data lands in **AI Center > Code Agents > Codex**, ChatGPT data in **AI Center > Code Agents > ChatGPT**, and the API Platform metrics in Metrics Explorer under the `openai_administration_` prefix.

## Where the raw records land

Besides metrics, the OpenAI sources write their raw records to predefined datasets in the `default` dataspace. Use them to go beyond the built-in AI Center dashboards: query a dataset with DataPrime to investigate activity record by record, and chart it in your own [Custom Dashboards](https://coralogix.com/docs/user-guides/custom-dashboards/introduction.md) or alerts. Each dataset page lists its fields and streams and includes a starter query. Grant access to these datasets too, so the Codex and ChatGPT dashboards can show the details that come from them:

| Source                                                                                                                                                                                                          | Dataset                                                                                                                                           | What it holds                                                                                                    |
| --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------- |
| OpenAI Administration APIs ([API Platform](https://coralogix.com/docs/integrations/ai-observability/openai/api-platform.md))                                                                                    | [`ai.raw.openai.admin_logs`](https://coralogix.com/docs/user-guides/data-layer/default-dataspace/ai-raw-openai-admin-logs.md)                     | Usage and cost rows, directory and governance objects, and audit events.                                         |
| ChatGPT Enterprise Compliance API ([ChatGPT](https://coralogix.com/docs/integrations/ai-observability/openai/chatgpt.md) and [Codex](https://coralogix.com/docs/integrations/ai-observability/openai/codex.md)) | [`ai.raw.openai.compliance_logs`](https://coralogix.com/docs/user-guides/data-layer/default-dataspace/ai-raw-openai-compliance-logs.md)           | Audit and sign-in events, conversation messages, the workspace inventory, Codex events and tasks, and cost rows. |
| Codex analytics per ChatGPT workspace ([Codex](https://coralogix.com/docs/integrations/ai-observability/openai/codex.md))                                                                                       | [`ai.raw.openai.codex.analytics_logs`](https://coralogix.com/docs/user-guides/data-layer/default-dataspace/ai-raw-openai-codex-analytics-logs.md) | Daily Codex usage per user and workspace, and Codex code review activity.                                        |

These datasets are restricted by default. To let more users query them, add a target group rule to the access policy of each dataset. See [Access control](https://coralogix.com/docs/user-guides/data-layer/dataset-management/access-control.md).