​
Getting it up and running

  1. Clone the repo from GitHub:
git clone https://github.com/Eigen-DB/eigen-db
  1. To start EigenDB, run:
docker compose up -d
Omit the -d flag if you don’t want to run the it in the background. Your EigenDB instance should now be running on port :8080 You can test the connection to your instance by running: 
curl http://localhost:8080/api/v1/health
  1. Get your API key either from apps/eigendb/eigen/api_key.txt or from the container logs:
docker logs CONTAINER_NAME

​
Creating your first index

EigenDB can be interfaced with through our Python SDK or its robust REST API.
To install the Python SDK, check out this page.
python
import os
from eigen_client.client import Client
from eigen_client.data_types import Document

client = Client(
    url="http://localhost:8080",
    api_key="eigendb-***",
)

index = client.create_index_from_model(
    index_name="food-facts",
    model_name="text-embedding-3-small",
    model_provider="openai",
    model_provider_api_key="your openai api key..."
)

documents = [
    Document(id=1, data="Fresh herbs boost flavor.", metadata={"recipe_id": "123"}),
    Document(id=2, data="Slow simmer blends soup.", metadata={"recipe_id": "456"}),
    Document(id=3, data="Homemade bread smells great.", metadata={"recipe_id": "789"}),
    Document(id=4, data="Grilled veggies taste sweeter.", metadata={"recipe_id": "987"}),
    Document(id=5, data="Cast iron sears steak well.", metadata={"recipe_id": "654"})
]

index.upsert_docs(documents)

results = index.search_docs(
    string="Baking",
    k=3
)

print(results)
To first create an index:
curl
curl -X PUT http://localhost:8080/api/v1/indexes/[INDEX-NAME]/create \
  -H "Content-Type: application/json" \
  -H "X-Eigen-API-Key: eigendb-***" \
  -d '{
    "dimensions": 1536,
    "metric": "cosine"
  }'
To upsert your embeddings:
curl
curl -X PUT http://localhost:8080/api/v1/embeddings/[INDEX-NAME]/upsert \ 
  -H "Content-Type: application/json" \
  -H "X-Eigen-API-Key: eigendb-***" \
  -d '{
    "embeddings": [
      {
        "id": 1,
        "data": [0.1, 0.2, 0.3, ..., 0.384],
        "metadata": {"recipe_id": "123"}
      },
      {
        "id": 2,
        "data": [0.2, 0.1, 0.4, ..., 0.385],
        "metadata": {"recipe_id": "456"}
      },
      ...
    ]
  }'
To search for similar embeddings:
curl
curl -X POST http://localhost:8080/api/v1/embeddings/[INDEX-NAME]/search \
  -H "Content-Type: application/json" \
  -H "X-Eigen-API-Key: eigendb-***" \
  -d '{
    "queryVector": [0.1, 0.2, 0.3, ..., 0.384],
    "k": 3
  }'