# TopK Documentation

> Hybrid search, multi-vector search, document parsing, question answering, and more in one API.

## Overview

- [Introduction](https://docs.topk.io/introduction)
- [Architecture](https://docs.topk.io/architecture)
- [Concepts](https://docs.topk.io/concepts)
- [Limits](https://docs.topk.io/limits): Production limits for collections and datasets.
- [Regions](https://docs.topk.io/regions)
- [Changelog](https://docs.topk.io/changelog)

## Dataset API

- [Ingest Files](https://docs.topk.io/datasets/ingest): Upload your document files to a dataset.
- [Ask](https://docs.topk.io/datasets/ask): Get grounded, evidence-backed answers from your documents.
- [Search](https://docs.topk.io/datasets/search): Retrieve the most relevant passages from your documents.

## Collection API

- [Write](https://docs.topk.io/collections/write): Upsert, update, or delete documents in a collection.
- [Query](https://docs.topk.io/collections/query): Query documents stored in a collection.
- [Get](https://docs.topk.io/collections/get): Get documents stored in a collection by `_id`.

## Management API

- [Datasets](https://docs.topk.io/datasets/manage): Create, list, get, update, and delete datasets.
- [Collections](https://docs.topk.io/collections/manage): Create, list, get, and delete collections.

## Guides

- [Semantic search](https://docs.topk.io/guides/semantic-search): Add state-of-the-art semantic search to your app with no embedding pipeline, no vector store, and no reranking service to maintain.
- [Multi-vector search](https://docs.topk.io/guides/multi-vector-search): Late-interaction retrieval using token-level embeddings and MaxSim scoring for long documents and multi-aspect queries.
- [Dense vector search](https://docs.topk.io/guides/vector-search): High-performance ANN search with >95% recall and p99 < 50ms latency for semantic similarity at scale.
- [Sparse vector search](https://docs.topk.io/guides/sparse-vector-search): Exact retrieval over high-dimensional sparse representations with 100% recall and support for learned sparse models like SPLADE.
- [Keyword search (BM25)](https://docs.topk.io/guides/keyword-search): Full-text keyword search using the BM25 ranking function for term-frequency-based document retrieval.
- [True Hybrid search](https://docs.topk.io/guides/true-hybrid-search): Combine vector, multi-vector, keyword search, and metadata filtering in a single query for best-of-all-worlds retrieval.

## Python SDK Reference

- [topk_sdk](https://docs.topk.io/sdk/topk-py)
- [topk_sdk.data](https://docs.topk.io/sdk/topk-py/data)
- [topk_sdk.error](https://docs.topk.io/sdk/topk-py/error)
- [topk_sdk.query](https://docs.topk.io/sdk/topk-py/query)
- [topk_sdk.schema](https://docs.topk.io/sdk/topk-py/schema)

## JavaScript SDK Reference

- [topk-js](https://docs.topk.io/sdk/topk-js)
- [topk-js/data](https://docs.topk.io/sdk/topk-js/Namespace.data)
- [topk-js/query_fn](https://docs.topk.io/sdk/topk-js/Namespace.query_fn)
- [topk-js/query_agg](https://docs.topk.io/sdk/topk-js/Namespace.query_agg)
- [topk-js/query](https://docs.topk.io/sdk/topk-js/Namespace.query)
- [topk-js/schema](https://docs.topk.io/sdk/topk-js/Namespace.schema)

## SQL

- [Overview](https://docs.topk.io/sdk/topk-sql/overview): Connect any PostgreSQL client to TopK and use SQL to create collections, insert data, and run search.