Bixel

TopK

Topk.io
Product Capabilities

Multi-tenancy platform

As of July 20, 2026, TopK's features: Benchmarked at 124.8ms P99 latency for multi-vector queries, 176.5 QPS per replica set, and a 1.2-second index lag, with 95% recall on filtered queries; those numbers are stated by the company and no independent benchmark is cited, but they give a concrete baseline to test against.

High confidence · 33 dated facts
Corroborated across multiple dated, sourced facts.
Capabilities
NamespacesMulti TenancyVector SearchSparse VectorsMetadata FilteringText EmbeddingFile SearchDocument ProcessingSemantic SearchMulti Vector EmbeddingAgentic QueriesCitationsHybrid SearchKeyword SearchCustom RerankingDense VectorsMulti Vector RetrievalRerankingOnline WritesHorizontal ScalingRead ReplicasSQL InterfaceLate InteractionManaged EmbeddingsAnn Search
Integrations
Postgres
Details
Ingest Throughput70 mb_s_per_partition
Vectors Per Second70000 vectors_s_per_partition
Max Items Per Partition1000000000 items
Multi Vector P99 Latency124.8 ms
Multi Vector Qps176.5 qps_per_replica_set
Index Lag1.2 s
Filter Benchmark Recall95
Features Change Historydated events · values unlock with a key
Feature added: Custom Reranking
Feature added: Vector Search
Feature added: Dense Vectors
Feature added: Horizontal Scaling
Feature added: Hybrid Search
Feature added: Postgres
Feature added: Keyword Search
Feature added: Late Interaction
Every features fact on this pagekey · value · provenance · dated · sourced
FactValueProvenanceAs ofSource
features.agentic-queriesAgentic Queriescompany stated2026-07-20www.topk.io/pricing
features.ann-searchAnn Searchcompany stated2026-07-20www.topk.io/use-cases/recsys
features.citationsCitationscompany stated2026-07-20www.topk.io/pricing
features.custom-rerankingCustom Rerankingcompany stated2026-07-18www.topk.io/blog/seed-round
features.dense-vectorsDense Vectorscompany stated2026-07-20www.topk.io/use-cases/recsys
features.document-processingDocument Processingcompany stated2026-07-20www.topk.io/pricing
features.file-searchFile Searchcompany stated2026-07-20www.topk.io/pricing
features.horizontal-scalingHorizontal Scalingcompany stated2026-07-20www.topk.io/use-cases/recsys
features.hybrid-searchHybrid Searchcompany stated2026-07-18www.topk.io/blog/seed-round
features.integration.postgresPostgrescompany stated2026-07-20www.topk.io/use-cases/recsys
features.keyword-searchKeyword Searchcompany stated2026-07-18www.topk.io/blog/seed-round
features.late-interactionLate Interactioncompany stated2026-07-20www.topk.io/use-cases/recsys
features.limit.filter-benchmark-recall95company stated2026-07-20www.topk.io/use-cases/recsys
features.limit.index-lag1.2 scompany stated2026-07-20www.topk.io/use-cases/recsys
features.limit.ingest-throughput70 mb s per partitioncompany stated2026-07-20www.topk.io/use-cases/recsys
features.limit.max-items-per-partition1000000000 itemscompany stated2026-07-20www.topk.io/use-cases/recsys
features.limit.multi-vector-p99-latency124.8 mscompany stated2026-07-20www.topk.io/use-cases/recsys
features.limit.multi-vector-qps176.5 qps per replica setcompany stated2026-07-20www.topk.io/use-cases/recsys
features.limit.vectors-per-second70000 vectors s per partitioncompany stated2026-07-20www.topk.io/use-cases/recsys
features.managed-embeddingsManaged Embeddingscompany stated2026-07-20www.topk.io/use-cases/recsys
features.metadata-filteringMetadata Filteringcompany stated2026-07-20www.topk.io/pricing
features.multi-tenancyMulti Tenancycompany stated2026-07-20www.topk.io/pricing
features.multi-vector-embeddingMulti Vector Embeddingcompany stated2026-07-20www.topk.io/pricing
features.multi-vector-retrievalMulti Vector Retrievalcompany stated2026-07-20www.topk.io/use-cases/recsys
features.namespacesNamespacescompany stated2026-07-20www.topk.io/pricing
features.online-writesOnline Writescompany stated2026-07-20www.topk.io/use-cases/recsys
features.read-replicasRead Replicascompany stated2026-07-20www.topk.io/use-cases/recsys
features.rerankingRerankingcompany stated2026-07-20www.topk.io/use-cases/recsys
features.semantic-searchSemantic Searchcompany stated2026-07-20www.topk.io/pricing
features.sparse-vectorsSparse Vectorscompany stated2026-07-20www.topk.io/pricing
features.sql-interfaceSQL Interfacecompany stated2026-07-20www.topk.io/use-cases/recsys
features.text-embeddingText Embeddingcompany stated2026-07-20www.topk.io/pricing
features.vector-searchVector Searchcompany stated2026-07-20www.topk.io/pricing
Features across Vector DatabasesTopK ranked in place · tap through for each read
CompanyFeatures readFacts
ZillizHybrid search platform79
QdrantHybrid search platform50
VectaraRAG platform48
ChromaHybrid search platform46
LanceDBHybrid search platform38
ActiveloopRAG platform37
TopK this recordMulti-tenancy platform33
PineconeReranking platform31
TurbopufferMulti-tenancy platform31
KDB.AIFocused feature set18
EpsillaRAG platform16
VespaHybrid search platform14
MarqoFocused feature set5

No observed features facts yet for Milvus, MyScale and Weaviate.

Get features for topk.io via API / MCPevery field dated and sourced
RESTopen tier
GET https://api.bixel.com/v1/companies/topk.io/facts?dimension=features

{
  "data": {
    "facts": [
      {
        "key": "features.namespaces",
        "value": true,
        "provenance": "company_stated",
        "as_of": "2026-07-20",
        "source_url": "www.topk.io/pricing"
      },
      {
        "key": "features.multi-tenancy",
        "value": true,
        "provenance": "company_stated",
        "as_of": "2026-07-20",
        "source_url": "www.topk.io/pricing"
      },
      {
        "key": "features.vector-search",
        "value": true,
        "provenance": "company_stated",
        "as_of": "2026-07-20",
        "source_url": "www.topk.io/pricing"
      },
      {
        "key": "features.sparse-vectors",
        "value": true,
        "provenance": "company_stated",
        "as_of": "2026-07-20",
        "source_url": "www.topk.io/pricing"
      },
      "…"
    ]
  }
}
MCPfor agents
# any MCP client (Claude, agents)
const record = await bixel.get_company_facts({ domain: "topk.io", dimension: "features" })

# returns the features record above,
# each value with its source_url + as_of,
# ready to reason over
Build on the company record. One key, REST + MCP, every signal dated and sourced back to the page it came from.

Public record, read from companies' own pages and boards. Every fact dated and sourced; provenance (observed vs company stated) shown inline.