Qdrant is an open-source vector database and search engine, made specifically for storing embeddings (numerical representations of images, text, audio, etc.) and searching for the closest matches by meaning and not just keywords. In practice, this enables AI-applications to quickly generate semantic search and recommendations by indexing regular data as multidimensional vectors with ANN algorithms.
Qdrant stores data as “collections” – named groups of “points”. In this case, a “point” is a vector (embedding), combined with optional ID and JSON-formatted metadata. You are given 2 options to store them:
- Memory storage – enables the highest performance by storing vectors directly in system memory, writing to disk in the background.
- Memmap – offers a balance between performance and storage efficiency by utilizing a memory mapping to a disk.
Qdrant developers provide ready-to-use client SDKs for:
And if your language of choice is not in this list, you can generate a client library based on either OpenAPI or protobuf specification.
Qdrant from Beget Cloud is a ready-to-use vector database and search engine installed on your virtualized server with a domain of your choice, secured with a free TLS certificate.
Includes
- Ubuntu
- Docker
- Qdrant
Qdrant installation
When creating a server, along with the regular parameters, you can specify:
- Domain name – for your Qdrant install, you can either bring your own domain, or use a free generated domain name in .beget.app zone.
- Administrator email – used for issuing a TLS certificate.
- API key – used to authorize the requests sent to Qdrant. Auto-generated by default.
Getting started with Qdrant
To begin using Qdrant, follow the address https://my-domain.beget.app:6333/dashboard, where my-domain.beget.app is the domain name selected during the installation process. You will be prompted to enter your API key:
To log in, enter the API key specified during installation. You can also find it in the “About Qdrant” section in the control panel. After logging in, you will be greeted by the dashboard:
Additional setup is not required – the service can be used right out of the gate. If you haven’t worked with Qdrant before, press “Quickstart” for a quick built-in tutorial. You can find more information on working with the API in the corresponding section of the Qdrant website, as well as detailed documentation.
FAQ
Qdrant is installed as a Docker container, you can find docker-compose.yml, as well as config.yaml and .env configuration files in the /opt/beget/qdrant directory. Directories storage and snapshots, used for persistent storage, are also located here.
We automatically install the latest version of Qdrant when creating a server. To update later, change to the app’s configuration directory with a cd /opt/beget/qdrant command, download the latest Docker image with docker compose pull and recreate the container with docker compose down and docker compose up -d.
To prevent unauthorized access to your Qdrant instance in case of a compromised API key, please follow these steps to update your API key:
- Connect to your server via SSH and change to the app’s configuration directory via
cd /opt/beget/qdrantcommand. - Open the
.envfile with a text editor of your choice, e.g.,vimand change the API key specified in theQDRANT_API_KEYvariable to a new random key and save the changes. - Recreate the container by issuing the
docker compose downanddocker compose up -dcommands.
To connect your n8n server to Qdrant, download the official Qdrant node in the workflow editor:
Add the desired action to a workflow and create a new credential in its settings:
In the following window, enter your Qdrant API key and server address specified during the installation:
