An overview of Weaviate publications on third-party platforms
Interviews with Weaviate community and machine learning experts
Keep up to date with Weaviate's content
Weaviate vector search engine
Guide for contributions to Weaviate
Weaviate Cloud Console login
Source code on Github
Join our community on Slack
Start for free and pay as you go per vector dimension stored and queried
All paid plans deliver unlimited capacity over three different tiers, so your DBs may scale seamlessly
Starting from $0.05 per 1 million vector dimensions
Signup and get to your first 10k data objects for free. No credit card necessary.
Starting from$0.00000005 per ML-vector dimension.
Choose from the three different SLA types, for startups and enterprises
Our pricing is built around vector dimensions stored and queried, and different SLA-tiers have different prices per dimension. The exact calculation can be found in the FAQ below.(not inclusive of discounts and taxes).
get a link to this price
Our pricing is designed to give you all the capabilities to build and test your applications for free. When you are ready to move to production, simply pick a plan that best suits your needs.
Let us help answer the most common questions you might have.
In Weaviate, you can attach a vector embedding to a data object. A vector embedding can have an x-amount of dimensions. Some vectors have a lot of embeddings (sometimes more than 10k), some just a few (e.g., 90). The more vector dimensions you store, the more infrastructure is needed to optimize and maintain performance, this is the reason why we calculate with individual dimensions. We believe it's the fairest and most accurate price to give you the best experience.
You pay for the total amount of embedding dimensions stored per data object and per data object queried. For example, if you have a 100-dimensional embedding and you store 1k documents that you query 1k times per month. You pay for 200k dimensions.(stored objects + objects queried) * embedding dimension size * SLA tier price per dimension
(stored objects + objects queried) * embedding dimension size * SLA tier price per dimension
No, the Weaviate Cloud Service (WCS) is a different solution using the same code as Weaviate open-source. The difference is the WCS itself (i.e., SaaS) and different SLA types (opposed to the open-source BSD3 license).
Yes, please reach out to us at firstname.lastname@example.org.
Not yet; we are currently in the process of becoming SOC2 compliant. Feel free to email us for a status update on email@example.com.
Hibernation is a process where a cluster goes down (i.e., "hybernates") while retaining your data after a given period. This is ideal for research or development purposes; when the service endpoints are used again, the service comes back up with a short time delay.
If your SLA tier contains round-robin provisioning, the Weaviate Cloud Service will provision on Amazon Web Services (AWS), Google Cloud Platform (GCP), or Microsoft Azure where enough resources are available.
No, storage of data objects without vector embeddings is on us.
Weaviate modules based on inference APIs are automatically integrated in the cloud service. These modules currently include Hugging Face Inference and OpenAI's embeddings end-points.
We take the median number of vectors per month measured per hour.
Usage spikes are -almost- always a good sign! Your Weaviate-powered app or platform is being actively used, and we don't want your bill to be in the way of your success. Spikes are analyzed at the end of the month, and occasional ones are on us.
The company behind Weaviate is SeMI Technologies. They run the Cloud Service and maintain the open source software.
We are currently onboarding customers onto the Weaviate Cloud Services via the Private Beta program. Please leave your contact details below if you want to join the first wave of managed vector search users. After leaving your details, a representative will reach out to you within 24 hours to investigate if you qualify for the private beta program.
Copy this link to share this pricing configuration