OpenAI (Public Preview)

OpenAI is a research organization that aims to create advanced artificial intelligence (AI) systems and promote safe and beneficial AI development. They pursue this goal through a variety of research projects, including natural language processing, reinforcement learning, computer vision, and robotics. OpenAI also provides tools and resources to enable other researchers and developers to build and experiment with AI technology. For more about OpenAI, please refer to the OpenAI Documentation.

The Observe OpenAI app ingests token usage data so you can get insights into your OpenAI application. You can get answers to questions such as the following:

• How can I optimize my use of the OpenAI API to minimize costs while still achieving good results?

• What models are being used to generate responses?

The Token Usage Monitoring Home dashboard gives a high-level overview of the various OpenAI API tokens you have in use. You can quickly get insights into the cost of each token, the number of requests, and the number of tokens generated per request.

Figure 1 - The Token Usage Monitoring Home dashboard

If you filter down to a specific token, you can go to the Token Usage Instance dashboard. This dashboard gives you usage specific to the token you are interested in, including requests broken down by API (language, dall-e, and whisper), token usage by snapshot, and a more detailed breakdown of cost.

Figure 2 - The Token Usage Instance dashboard

To set up the application, use the following steps:

1. Install the OpenAI app in Observe.

2. Click on Create Poller.

3. Add the OpenAI token you want to monitor in the OpenAI Token field.

Figure 3 - Setting up the OpenAI Poller