# Parakeet MLX

## Docs

- [Beam Decoding](https://mintlify.wiki/senstella/parakeet-mlx/advanced/beam-decoding.md): Improve transcription accuracy with beam search decoding
- [Local Attention](https://mintlify.wiki/senstella/parakeet-mlx/advanced/local-attention.md): Optimize memory usage with local attention for long audio transcription
- [Low-Level API](https://mintlify.wiki/senstella/parakeet-mlx/advanced/low-level-api.md): Direct access to audio preprocessing and model internals
- [Sentence Splitting](https://mintlify.wiki/senstella/parakeet-mlx/advanced/sentence-splitting.md): Control how transcriptions are segmented into sentences
- [AlignedResult](https://mintlify.wiki/senstella/parakeet-mlx/api/aligned-result.md): Top-level transcription result with sentences and tokens
- [AlignedSentence](https://mintlify.wiki/senstella/parakeet-mlx/api/aligned-sentence.md): Sentence-level transcription with timing and tokens
- [AlignedToken](https://mintlify.wiki/senstella/parakeet-mlx/api/aligned-token.md): Token-level transcription with precise timing
- [Audio Utilities](https://mintlify.wiki/senstella/parakeet-mlx/api/audio-utils.md): Functions for loading and preprocessing audio
- [BaseParakeet](https://mintlify.wiki/senstella/parakeet-mlx/api/base-parakeet.md): Base class for all Parakeet ASR models
- [DecodingConfig](https://mintlify.wiki/senstella/parakeet-mlx/api/decoding-config.md): Configuration for decoding behavior and sentence splitting
- [from_pretrained](https://mintlify.wiki/senstella/parakeet-mlx/api/from-pretrained.md): Load a Parakeet model from Hugging Face Hub or local directory
- [ParakeetCTC](https://mintlify.wiki/senstella/parakeet-mlx/api/parakeet-ctc.md): Connectionist Temporal Classification model implementation
- [ParakeetRNNT](https://mintlify.wiki/senstella/parakeet-mlx/api/parakeet-rnnt.md): RNN-Transducer model implementation
- [ParakeetTDT](https://mintlify.wiki/senstella/parakeet-mlx/api/parakeet-tdt.md): Token-and-Duration Transducer model implementation
- [ParakeetTDTCTC](https://mintlify.wiki/senstella/parakeet-mlx/api/parakeet-tdtctc.md): Hybrid TDT-CTC model combining TDT and CTC decoders
- [SentenceConfig](https://mintlify.wiki/senstella/parakeet-mlx/api/sentence-config.md): Configuration for splitting transcription into sentences
- [Decoding Strategies](https://mintlify.wiki/senstella/parakeet-mlx/concepts/decoding.md): Understanding greedy and beam search decoding algorithms in Parakeet MLX
- [Model Architectures](https://mintlify.wiki/senstella/parakeet-mlx/concepts/models.md): Understanding Parakeet MLX's ASR model variants and their architectural components
- [Timestamps & Alignment](https://mintlify.wiki/senstella/parakeet-mlx/concepts/timestamps.md): Understanding how Parakeet MLX computes word-level timestamps and aligns text to audio
- [Chunking Long Audio](https://mintlify.wiki/senstella/parakeet-mlx/guides/chunking.md): Efficiently process long audio files using chunking with overlap
- [CLI Usage](https://mintlify.wiki/senstella/parakeet-mlx/guides/cli-usage.md): Complete guide to using the Parakeet MLX command-line interface
- [Output Formats](https://mintlify.wiki/senstella/parakeet-mlx/guides/output-formats.md): Export transcriptions in TXT, SRT, VTT, and JSON formats
- [Python API](https://mintlify.wiki/senstella/parakeet-mlx/guides/python-api.md): Complete guide to using Parakeet MLX in your Python applications
- [Streaming Transcription](https://mintlify.wiki/senstella/parakeet-mlx/guides/streaming.md): Real-time audio transcription with streaming inference
- [Installation](https://mintlify.wiki/senstella/parakeet-mlx/installation.md): Install Parakeet MLX using pip, uv, or as a standalone CLI tool
- [Introduction to Parakeet MLX](https://mintlify.wiki/senstella/parakeet-mlx/introduction.md): High-performance automatic speech recognition for Apple Silicon using Nvidia's Parakeet models
- [Quickstart](https://mintlify.wiki/senstella/parakeet-mlx/quickstart.md): Get started with Parakeet MLX in under 2 minutes