The retrieval engine beyond

Hybrid search, multi-vector retrieval, and custom ranking in one query. High-quality search built on object storage for 10x lower cost and massive scale.

1B+ docs

per collection

17ms p99

query latency

70MB/s

write throughput

Up to80%

higher recall

See full benchmarks →
topk_search.py
topk.collection("earnings_reports").query(
select(
"_id", "title",
# Semantic similarity
semantic_score=fn.semantic_similarity(
"content",
"NVDA data center revenue in Q4 2025",
),
# Multi-vector retrieval
visual_score=fn.multi_vector_distance(
"page_embedding",
[[0.97, 0.17, ..], [0.14, 0.99, ..], ..],
),
)
# Keyword search
.filter(match("nvidia") | match("nvda"))
# Metadata filtering
.filter(field("fiscal_year").eq(2025))
# Custom scoring
.sort(
(field("semantic_score") * 0.7 + field("visual_score") * 0.3)
* field("source_quality"),
asc=False,
)
.limit(10)
)
TopK Platform

Your search pipeline. In one API.

TopK is a unified retrieval engine for search, RAG, and agents. Combine semantic, dense/sparse vector, late interaction, and lexical search with metadata filters and custom scoring in a single query. No complex pipeline to build and manage.

Inference

EmbeddingOCRDocument Parsing

Unified Retrieval

Vector SearchLate InteractionKeyword SearchCustom Scoring
Agent
Human
Benchmarks

Higher accuracy. Faster. Cheaper.

TopK delivers state-of-the-art retrieval quality and database performance enabling more accurate answers while burning up to 10x fewer tokens.

Query latency

p99 · 1M vectors
Security & compliance

Security & Data protection

TopK is built from the ground up with enterprise security in mind. Data is encrypted in transit and at rest, access is scoped by role, and our infrastructure is audited continuously. When you need full control, we can deploy to your VPC or on-prem.

Encryption at rest & in transit
Role-based access control
Audit logging
Private deployment
SOC 2 Type I certified
SOC 2 Type IView Trust Center
FAQs

Common questions

Ship better search today.

Start building for free. Move to production with usage-based pricing or private deployment in your VPC.