Add AI search to existing application

Adding semantic search to an existing application involves three straightforward steps. First, add a vector column to the table you want to search. In a PostgreSQL database, you can utilize the pgvector extension, which adds a vector type and distance operators for efficient ranking. Create the column using the following SQL command:

```

CREATE EXTENSION IF NOT EXISTS vector;

ALTER TABLE todos ADD COLUMN IF NOT EXISTS embedding vector(1536);

```

Next, convert the text of each row into a 1536-number embedding using an embedding model. This is achieved by sending the text to the embedding model via an HTTP POST request using an API key. The model returns a list of numbers representing the semantic meaning of the text. Store these numbers in the newly created vector column.

Finally, when searching for a term, perform the same embedding process on the search query and query the database for the closest matching rows. This is done by calculating the distance between the search term's embedding and the embeddings stored in the table. In Postgres, this can be achieved using the built-in distance operators on the vector column.

With these three changes, you can add AI-powered semantic search to your existing application, allowing users to find relevant items based on the meaning of their queries rather than just matching substrings.