TopK

What it does

TopK is a retrieval stack that combines an engine, models and inference. It runs on object storage and supports dense and sparse vectors, multi-vector retrieval, regex and metadata filters, custom ranking, and managed inference through one API. Teams can sign up, get an API key, create collections, upsert documents and query indexed data using a Python SDK, a JavaScript SDK or SQL. Pricing, benchmarks, docs and a console are published on the official site.

The documentation lists Python and JavaScript SDKs plus a command line interface, with regions available for clients. Security notes state that data is encrypted in transit and at rest, with role based access control and auditability, and that private deployment inside a customer VPC is offered. TopK states it is SOC 2 Type I certified. The site describes use cases for enterprise search, visual document retrieval and agentic search. Collection, write, query and management APIs are documented. No pricing amounts are given here beyond the model cost mentioned in the release note on the home page.