Predictable pricing.

Pay only for what you use. Transparent, usage-based pricing, billed monthly.

YOUR WORKLOAD
10M
100K10B
1M
1K10B
1M
1K10B
10
110M
ESTIMATED MONTHLY COST
$15.79/mo
Built to scale. Priced to make sense.
Storage34.57 GiB$3.46
Write requests10K requests$0.10
Write data2.38 GiB$1.19
Query memory501 MiB$9.79
Query compute & requests1M queries$1.26
Start with $10 free credits

Indicative estimate in USD. Assumes 100 documents per write request. Egress, get operations, and embedding usage are billed separately.

PRICING BREAKDOWN

Units you understand.

A straightforward model for the entire retrieval stack.
All prices in USD.

ResourceUnitPrice
StorageVector index and metadataper GiB / month$0.10
Write requestsBatched upsert operationsper 1K requests$0.01
Write dataVector and metadata transferredper GiB written$0.50
Query requestsRequests against the indexper 1M queries$0.50
Query memoryHot memory serving your queriesper GiB / month$20.00
Query computeData scanned across the indexper TiB scanned$0.005
Data egressData returned from queries and getsper GiB returned$0.20
Get operationsFull document retrievalper 1K gets$0.01
Text embeddingBuilt-in multi-vector embeddingsper 1M tokens$0.05

YOUR INFRASTRUCTURE. OUR ENGINE.

Need a private deployment?

Bring TopK to your VPC or on-prem infrastructure.
Let’s find the right setup for your team.

Talk to us

Ship better
search today.

Build on flexible primitives and focus on
what makes your beer taste better.

Read the docs