chonkie Releases
120 releases of feyninc/chonkie
- v1.2.0
A major feature update introducing a new advanced SemanticChunker and support for multiple vector database handshakes.
Aug 28, 2025
- v1.1.2
Adds support for new embedding providers and utility classes for file handling and text preparation.
Aug 6, 2025
- v1.1.1
Resolves Windows installation issues and introduces a new dataset porting utility.
Jul 18, 2025
- v1.1.0
Focuses on documentation updates, static typing improvements, and simplifying the tokenizer configuration.
Jul 2, 2025
- v1.0.10
Enhances type safety by adding mypy stubs and ensuring the package is properly typed for IDE support.
Jun 7, 2025
- v1.0.9
Performance optimizations via Cython and the addition of support for PostgreSQL vector databases.
Jun 4, 2025
- v1.0.8
Expands API compatibility and adds full support for cloud chunkers and URI-based model selection.
May 22, 2025
- v1.0.7
Introduces core infrastructure for data storage (Handshakes/Porters), new embedding providers, and visualizer themes.
May 6, 2025
- v1.0.6
This release introduces advanced chunking strategies including the agentic `SlumberChunker` and neural `NeuralChunker`, alongside the new `Genie` framework for API integration.
Apr 28, 2025
- v1.0.5
A minor patch release focused on fixing import issues to ensure `CodeChunker` is properly accessible from the main module.
Apr 22, 2025
- v1.0.4
Added support for code-specific chunking, Jina AI embeddings, and new refineries to manage chunk overlap and embeddings.
Apr 21, 2025
- v1.0.3
Introduced a visualization tool for debugging chunk quality and a recipe system for multilingual and document-specific chunking.
Apr 14, 2025
- v1.0.2
Bug fixes for serialization and refactoring of the `RecursiveChunker`.
Apr 7, 2025
- v1.0.1
Version bump to resolve PyPI release issues.
Apr 6, 2025
- v1.0.0
This is the initial stable release (v1.0.0) of Chonkie, featuring bug fixes for the `__dict__` attribute and cleanup of unused `LateSentence` code, alongside version metadata updates.
Apr 6, 2025
- v3.1.2GUI for precise control
Introduces a new GUI for fine-grained control over speech synthesis, allowing users to adjust pitch and duration and swap voices across 7000 languages.
Oct 7, 2024
- v3.1.1Improved Vocoder through end-to-end Training
Enhances audio quality by retraining the vocoder end-to-end on synthetic spectrograms and increasing its capacity, resulting in a significant quality boost without sacrificing inference speed.
Sep 22, 2024
- v3.1Improved TTS in 7000 Languages
Builds upon v3.0 with new checkpoints and stochastic prosody prediction to handle 7000 languages more effectively, along with expanded IPA support and overhauled language similarity modules.
Jul 25, 2024
- v3.0TTS in 7000 Languages
Launches a massively multilingual TTS model supporting over 7000 languages through extrapolation from a 462-language checkpoint, featuring improved quality, audio watermarking, and various quality-of-life improvements as presented at Interspeech 2024.
Jun 10, 2024
- 2.pPrompting Controlled Emotional TTS
Adds functionality to condition TTS synthesis on emotional prompts, enabling the transfer of specific emotions to synthesized speech based on user-provided text during inference.
Jun 10, 2024