# OpenAI

On Epsilla Cloud, you can enable OpenAI integration by providing your OpenAI API key (we securely manage your keys using AWS KMS):

<figure><img src="https://2532879721-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM0ZX7fId7ifK45ldHWEp%2Fuploads%2Fgit-blob-429d6b5ac4d929ef2acd843f8c254d7793cd25aa%2FScreenshot%202024-05-18%20at%208.51.21%20AM.png?alt=media" alt=""><figcaption></figcaption></figure>

## Embeddings

Epsilla integrates with OpenAI with the following embedding models:

<table><thead><tr><th width="311">Name</th><th width="171">Dimensions</th><th>Support Dimension Reduction</th></tr></thead><tbody><tr><td><strong>openai/text-embedding-3-large</strong></td><td>3072</td><td>Yes</td></tr><tr><td><strong>openai/text-embedding-3-small</strong></td><td>1536</td><td>Yes</td></tr><tr><td><strong>openai/text-embedding-ada-002</strong></td><td>1536</td><td>No</td></tr></tbody></table>

For Epsilla open source vector db, you just need to add a header in the data ingestion and semantic search queries [like this](https://epsilla-inc.gitbook.io/epsilladb/advanced-topics/embeddings#openai-embedding).

Then you can start using the openai embedding model during vector table schema creation:

<figure><img src="https://2532879721-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM0ZX7fId7ifK45ldHWEp%2Fuploads%2Fgit-blob-fe1a2ae2c43d3304299cf1efd978370dbbba2f1e%2FScreenshot%202024-01-31%20at%2012.10.57%20PM.png?alt=media" alt=""><figcaption></figcaption></figure>

For models that support dimension reduction, you can optionally provide a dimensions parameter to reduce the embedding response size:

<figure><img src="https://2532879721-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FM0ZX7fId7ifK45ldHWEp%2Fuploads%2Fgit-blob-b3a8bd1e3731255ba407e84aef85eaf737876776%2FScreenshot%202024-02-13%20at%205.29.51%20PM.png?alt=media" alt=""><figcaption></figcaption></figure>