Vekraris — Cloud Vector Database on ICP
Overview
Vekraris is a next-generation cloud vector database built on the Internet Computer Protocol (ICP). Designed for high performance, scalability, and security, Vekraris lets you store, search, and manage embeddings and high-dimensional data efficiently for AI, search, and data analytics applications.
Endpoints & Example Usage
1. Hybrid Search
Performs a hybrid (vector + keyword) search on your dataset.
Request:
curl --location 'http://localhost:3000/hybridSearch' \
--header 'Content-Type: application/json' \
--data-raw '{
"userKey": "1234567",
"dbName": "OmegaData",
"account": "test@example.com",
"query": "multi-factor authentication",
"limit": 1
}'
Response:
[
{
"title": "System Scalability Test Results",
"content": "Scalability tests show system handles 10,000 concurrent users with 95% uptime. Latency spikes observed at peak load. Recommended adding two more nodes to cluster for improved resilience.",
"createdAt": "1752489221237955988",
"score": 0.08027743473736361
}
]
2. Get Embeddings
Retrieves stored embeddings from your database.
Request:
curl --location 'http://localhost:3000/getEmbeddings' \
--header 'Content-Type: application/json' \
--data-raw '{
"userKey": "1234567",
"dbName": "OmegaData",
"account": "test@example.com",
"start": 0,
"limit": 200
}'
Response: Returns a list of embedding objects associated with your data. (Response schema varies by implementation.)
3. Query Embeddings
Finds the most relevant embeddings for a text query.
Request:
curl --location 'http://localhost:3000/queryEmbeddings' \
--header 'Content-Type: application/json' \
--data-raw '{
"userKey": "1234567",
"dbName": "OmegaData",
"account": "test@example.com",
"query": "incremental backups"
}'
Response:
[
{
"title": "System Scalability Test Results",
"content": "Scalability tests show system handles 10,000 concurrent users with 95% uptime. Latency spikes observed at peak load. Recommended adding two more nodes to cluster for improved resilience.",
"createdAt": "1752489221237955988",
"similarity": 0.22472182796853496
}
]
Key Features
- Cloud-Native: Built for scalability and high availability on ICP.
- Hybrid Search: Combines keyword and vector similarity for more relevant results.
- Fast Retrieval: Optimized for rapid search across large vector datasets.
- Secure: User and database-level API key authentication.
- Rich Metadata: Store and retrieve custom metadata with every vector.
Tips
- Replace the endpoint URL with your live API address when deploying.
- Always keep your
userKey
and account information confidential. - Fine-tune the
limit
parameter for your application's performance needs. - Use descriptive
dbName
values for organizing multiple vector databases.