No fuss vector database for AI, full documentation available at SemaDB repository.

Why SemaDB?

No fuss means: No pod size calculations, no schema definitions, no partition settings, no parameter tuning, no search algorithm tuning, no complex installation, no complex API.

• Hosted via RapidAPI: Quickly integrate with an easy-to-use vector database with RapidAPI and a single account. Focus on building the AI tool instead of worrying about the internals of a vector database.
• Lower cost: SemaDB offers one of the lowest possible costs for using a hosted vector database compared to other alternatives. You can get started for free and then make the most of the transparent pricing on RapidAPI.
• High recall: SemaDB achieves high recall across some standard vector similarity search benchmarks. The search results are optimised to return similar points as efficiently as possible.
• Self-contained: SemaDB is self-contained, built from the ground up and doesn’t rely on third party services.
• Automatic sharding: Store millions of points with automatic sharding that adapts to your data size. You don’t have to worry about partitioning, distributing or tuning your database.
• Multiple collections: Organise your data into multiple collections with ease and ensure the search is performed within a boundary.

Get Started in 3 Steps

Getting started with SemaDB only takes 3 API calls:

1. Create a collection using the /collections API such as products.
2. Insert points into the collection via /collections/{collectionId}
3. Search for similar points using /collections/{collectionId}/search

Your account details are automatically integrated with RapidAPI and are used by SemaDB.

What are vector databases?

A vector database is a specialized database designed to store and manage vector data. Vector data is data that is represented as a multidimensional point in a vector space. This type of data is often used in machine learning and artificial intelligence applications, such as natural language processing, computer vision, and recommendation systems.

Vector databases offer a number of advantages over traditional relational databases for storing and managing vector data. These advantages include:

• Efficiency: Vector databases are optimized for efficient storage and retrieval of vector data. This makes them ideal for applications that require fast similarity searches, such as image search and product recommendations.
• Scalability: Vector databases can scale horizontally to handle large volumes of data. This makes them suitable for enterprise-grade applications.
• Ease of use: Vector databases provide user-friendly interfaces and APIs for managing vector data. This makes them easy to use for developers and data scientists.

Use Cases

As a vector database, SemaDB can be used in a wide variety of applications, including:

• Semantic text search: Instead of relying on keywords, you can use neural network based embedding models to search through semantically similar words, sentences or whole documents. These often are used in large language models to generate contextual responses.
• Image and video search: Vector databases can be used to power image and video search engines. For example, a vector database can be used to find images that are similar to a given image, even if the images are visually different.
• Recommendation systems: Vector databases can be used to power recommendation systems. For example, an e-commerce website can use a vector database to recommend products to users based on their past purchases and browsing behaviour.
• Natural language processing: Vector databases can be used to store and manage word embeddings, which are used in a variety of natural language processing tasks, such as machine translation and text summarization.
• Fraud detection: Vector databases can be used to detect fraudulent transactions and activities. For example, a bank can use a vector database to compare the features of a transaction to the features of known fraudulent transactions.
• Anomaly detection: Vector databases can be used to detect anomalies in data. For example, a manufacturing company can use a vector database to detect anomalies in the production process.
• Product discovery: Vector databases can be used to help users discover new products. For example, an e-commerce website can use a vector database to recommend products to users based on their past purchases and browsing behaviour.
• Personalization: Vector databases can be used to personalise user experiences. For example, a social media platform can use a vector database to recommend content to users based on their interests.

Limits

The API limits are documented in the API schema. In summary, the following limits are important to keep in mind:

• Collection IDs have a maximum length of 24 characters.
• Maximum request size is 20MB regardless of plan. This is set at the RapidAPI proxy level. Larger requests will be rejected.
• Maximum number of points per insertion request is 10k but this may depend on the dimensions of the vector and the metadata. It will be capped by the request size limit above. Large point insert requests may take longer.
• Maximum vector size / dimensions is 4096.
• Maximum number of points per update or deletion request is 100.
• Maximum search limit is 75 per request.