> ## Documentation Index
> Fetch the complete documentation index at: https://portkey-docs-add-third-party-integration-issues-fixes.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Supabase

Supabase provides an open source toolkit for developing AI applications using Postgres and pgvector. With Portkey integration, you can seamlessly generate embeddings using AI models like OpenAI and store them in Supabase, enabling efficient data retrieval. Portkey’s unified API supports over 250 models, making AI management more streamlined and secure

## Prerequisites

1. Supabase project API Key
2. [Portkey](https://app.portkey.ai/?utm_source=supabase\&utm_medium=content\&utm_campaign=external) API key

## Setting up your environment

First, let's set up our Python environment with the necessary libraries:

```sh theme={null}
pip install portkey-ai supabase
```

## Preparing your database

<Steps>
  <Step title="Create a Supabase Project">
    Go to [Supabase](https://supabase.com/dashboard/sign-in) and create a new project.
  </Step>

  <Step title="Enable pgvector extension">
    `pgvector` is an extension for PostgreSQL that allows you to both store and query vector embeddings within your database. We can enable it from the web portal through Database → Extensions. You can also do this in SQL by running:

    ```sql theme={null}
    create extension vector;
    ```
  </Step>

  <Step title="Create a table to store our documents and their embeddings">
    `pgvector` introduces a new data type called `vector`. In the code above, we create a column named `embedding` with the `vector` data type. The size of the vector defines how many dimensions the vector holds. OpenAI's `text-embedding-ada-002` model outputs 1536 dimensions, so we will use that for our vector size.
    We also create a `text` column named `content` to store the original document text that produced this embedding. Depending on your use case, you might just store a reference (URL or foreign key) to a document here instead.

    ```sql theme={null}
    create table documents (
      id bigserial primary key,
      content text,
      embedding vector(1536)
    );
    ```
  </Step>
</Steps>

## Configuring Supabase and Portkey

Now, let's import the required libraries and set up our Supabase and Portkey clients:

```python theme={null}
from portkey_ai import Portkey
from supabase import create_client, Client

# Supabase setup
supabase_url = "YOUR_SUPABASE_PROJECT_URL"
supabase_key = "YOUR_SUPABASE_API_KEY"
supabase: Client = create_client(supabase_url, supabase_key)

# Portkey setup
portkey_client = Portkey(
    api_key="YOUR_PORTKEY_API_KEY",
    provider="openai",
    virtual_key="YOUR_OPENAI_VIRTUAL_KEY",
)
```

Replace the placeholder values with your actual Supabase and Portkey credentials.

## Generating and storing embeddings

Let's create a function to generate embeddings using Portkey and OpenAI, and store them in Supabase:

```python theme={null}
#Generate Embedding
embedding_response = client.embeddings.create(
  model="text-embedding-ada-002",
  input="The food was delicious and the waiter...",
  encoding_format="float"
)

embedding = embedding_response.data[0].embedding

# Store in Supabase
result = supabase.table('documents').insert({
    "content": text,
    "embedding": embedding
}).execute()
```

This function takes a text input, generates an embedding using  through Portkey, and then stores both the original text and its embedding in the Supabase `documents` table.

Portkey supports 250+ Models, you can choose any model just by changing the `provider` and `virtual_key`

Here's an example on how to use `Cohere` with Portkey

```python theme={null}
client = Portkey(
    api_key="YOUR_PORTKEY_API_KEY", # defaults to os.environ.get("PORTKEY_API_KEY")
    provider="cohere",
    virtual_key="YOUR_COHERE_VIRTUAL_KEY",
)

embeddings = client.embeddings.create(
  model="embed-english-v3.0",
  input_type="search_query",
  input="The food was delicious and the waiter...",
  encoding_format="float"
)
```

<Info>
  Note that you will need to make a new table with `1024` dimensions instead of `1536` dimensions for Cohere's `embed-english-v3.0` model.
</Info>