Description
SvectorDB is a vector database built from the ground up for serverless environments, allowing developers to focus on their product rather than database management. It simplifies scaling from a single vector to millions, providing a high-performance and cost-effective solution that can be up to 20x cheaper than alternatives. The platform supports a quick start with intuitive JavaScript and Python SDKs, as well as an OpenAPI specification for integration with any language.
Read More
Key capabilities include hybrid search, enabling filtering of results based on key-value pairs alongside vector similarity. SvectorDB offers instant updates for upserts and deletions, eliminating concerns about eventual consistency. Its natively serverless architecture means pay-per-request pricing with no provisioning or scaling required. CloudFormation support allows for easy integration into existing infrastructure. Users can leverage built-in vectorizers for text and images or bring their own embeddings.
The pricing model is transparent and pay-per-request, with no minimum fees or upfront costs. Storage is priced at $0.25 per GB per month, and queries cost $5 per million operations. Writes are $20 per million operations. A free tier is available, offering up to 10 free-tier indexes with up to 5,000 records each, with no time limit.
SvectorDB is developed by a micro start-up, emphasizing transparency and direct customer interaction. While this offers hyper-responsiveness, potential customers should be aware of limitations such as the absence of database snapshots and a default maximum record limit of 1 million per database, which can be increased by contacting support. The company size is also a factor to consider for some clients.
Use cases span recommendation engines, document and image search, and retrieval-augmented generation for enhancing generative AI models. The platform supports embedding text and images using models like ALL_MINILM_L6_V2 and CLIP_VIT_BASE_PATH32, or custom embeddings.
- CategoryVector Databases & Retrieval
- Pricing ModelPay per request
- DeploymentServerless
SvectorDB's Core Features
Serverless architecture
Pay-per-request pricing
Hybrid search capabilities
Instant updates for data operations
Built-in vectorizers for text and images
Support for custom embeddings
OpenAPI specification for broad language support
CloudFormation integration
Scales from prototype to production
Cost-effective compared to alternatives
Free tier available
Transparent pricing model
How to use SvectorDB?
Configure: Set up your database ID and choose your preferred SDK (JavaScript, Python) or OpenAPI.
Embed Data: Use built-in vectorizers or your own embeddings to convert text or images into vectors.
Store Data: Use the `setItem` (or `set_item`) function to store items with their values and vectors.
Query Data: Perform vector similarity searches or key-based queries using the `query` function.
Filter Results: Utilize hybrid search to filter results by key-value pairs alongside vector similarity.
Scale: SvectorDB automatically handles scaling from small prototypes to millions of vectors.
Integrate: Connect SvectorDB into your existing applications and CloudFormation templates.
SvectorDB's Use Cases
- Recommendation Engines
- Document Search
- Image Search
- Retrieval Augmented Generation
- Prototyping AI Features
- Data Filtering and Analysis
