AI Vector Search

🧠 Native AI Vector Search (HNSW)

VortexKV allows storing high-dimensional vector embeddings and performing top-K nearest neighbor searches natively over standard Redis protocol without external plugins.

---

⚡ Vector Commands Reference

1. VADD <index> <id> <dim1> <dim2> ... <dimN>

Stores a high-dimensional vector in index <index> associated with <id>.

# Add document embeddings (4-dimensional example):
redis-cli -p 7379 -a "vortex_secure_2026" VADD articles doc_ai 0.95 0.05 0.0 0.0
redis-cli -p 7379 -a "vortex_secure_2026" VADD articles doc_quantum 0.1 0.9 0.0 0.0
redis-cli -p 7379 -a "vortex_secure_2026" VADD articles doc_crypto 0.2 0.3 0.9 0.0

2. VSEARCH <index> <topK> <metric> <q1> <q2> ... <qN>

Searches the top-K nearest neighbors to query vector [q1, ..., qN]. Supported metrics: cosine, euclidean, dot.

redis-cli -p 7379 -a "vortex_secure_2026" VSEARCH articles 2 cosine 0.90 0.10 0.0 0.0
# 1) 1) "doc_ai"
#    2) "0.999512"
# 2) 1) "doc_crypto"
#    2) "0.452110"

3. VSIM <index> <id1> <id2> <metric>

Calculates similarity score directly between two stored vectors.

redis-cli -p 7379 -a "vortex_secure_2026" VSIM articles doc_ai doc_quantum cosine
# "0.141421"

4. VDEL <index> <id>

Deletes a vector from the index.

redis-cli -p 7379 -a "vortex_secure_2026" VDEL articles doc_crypto
# (integer) 1

5. VINFO <index>

Retrieves stats and dimensions of a vector index.

redis-cli -p 7379 -a "vortex_secure_2026" VINFO articles
# 1) "index_name"
# 2) "articles"
# 3) "dimensions"
# 4) (integer) 4
# 5) "total_vectors"
# 6) (integer) 2