# MAX AI kernels API (Mojo)

> The MAX AI kernels library reference (top-level package indexes).

Version: Nightly

This file contains links to documentation sections following the llmstxt.org standard.

## Table of Contents

- [algorithm](https://docs.modular.com/api/kernels/algorithm/): Free-form row-wise reduction library: the `ReduceOp` monoid trait, the `rowwise` scaffolder (`Row` / `reduce` / `map` / `emit`), and the Row-based ...
- [builtin_kernels](https://docs.modular.com/api/kernels/builtin_kernels/): Provides the graph compiler's built-in kernel registrations.
- [comm](https://docs.modular.com/api/kernels/comm/): Provides communication primitives for multi-GPU workloads.
- [extensibility](https://docs.modular.com/api/kernels/extensibility/): Surface needed to write GraphCompiler kernel entry points.
- [kernels](https://docs.modular.com/api/kernels/): High-performance compute kernels, operators, and tensor APIs for MAX graphs.
- [kv_cache](https://docs.modular.com/api/kernels/kv_cache/): Contains implementations for several types of key-value caches.
- [linalg](https://docs.modular.com/api/kernels/linalg/): Provides CPU and GPU implementations of linear algebra functions.
- [nn](https://docs.modular.com/api/kernels/nn/): Provides neural network operators for deep learning models.
- [pipeline](https://docs.modular.com/api/kernels/pipeline/): Generic compile-time software pipelining for GPU kernels: schedule generation and verification.
- [profiling_range](https://docs.modular.com/api/kernels/profiling_range/): Provides Mojo wrappers for MAX's libkineto-backed range profiler.
- [quantization](https://docs.modular.com/api/kernels/quantization/): This package contains a set of APIs for quantizing tensor data.
- [shmem](https://docs.modular.com/api/kernels/shmem/): Implements a subset of OpenSHMEM for multi-node GPU communication.
- [state_space](https://docs.modular.com/api/kernels/state_space/): State space model kernels for Mamba and related architectures.
- [structured_kernels](https://docs.modular.com/api/kernels/structured_kernels/): Shared GPU kernel primitives for structured kernel architectures.