Appgithub-com-coffeemethod-kokorogui

Install Pinokio

Log in Register

Log in Register

KokoroGUI

https://github.com/CoffeeMethod/KokoroGUIupdated 2/21/2026, 1:11:04 AMindexed 3/1/2026, 1:59:38 AM

An advanced TTS software, built for audiobooks, podcasts, videos, and more.

Loading community details…

Community

Sort by

Post about KokoroGUI...Post

Loading...

Loading project activity…

Pinokio

KokoroGUI on Pinokio

Pinokio Apps Using This Repo

No Pinokio apps using this repo yet.

Community tagsNone yet.

Check-ins (0)

Platforms (0)

No reports yet.

Arch (0)

No reports yet.

GPU (0)

No reports yet.

RAM (0)

No reports yet.

VRAM (0)

No reports yet.

Recent commits

Merge pull request #2 from CoffeeMethod/dependabot/pip/pypdf-6.7.1

Coffee Method7 months ago4d32951

Bump pypdf from 6.6.2 to 6.7.1

dependabot[bot]7 months agoe25ba81

Update README with new screenshot and remove old image

Coffee Method7 months ago876899d

updated README.md for 3.1.0 update

CoffeeMethod7 months agodf4af1b

Added run.bat for easy startup on Windows.

CoffeeMethod7 months agofb03e07

1. Added `load_fx_preset(name)` to `KokoroEngine`: This method loads FX configurations from presets/fx/{name}.json. 2. Updated `parse_multispeaker_text(text)`: It now supports the [Speaker:FX]: syntax and returns a tuple of (speaker_name, fx_name, segment_text). 3. Updated `process_audio`: It now respects the apply_fx flag, allowing it to skip Pedalboard FX processing if disabled. 4. Updated engine methods: generate_preview, _process_jit_async, and _process_text_async now handle the new return signature of parse_multispeaker_text and correctly apply both speaker presets and the newly supported FX overrides.

CoffeeMethod7 months ago619892c

- UI Integration: Added a "JIT Generation" toggle in the Settings menu. When enabled, the "Start Generation" button transforms into "Start Real-time JIT". - Real-time Engine: Implemented start_jit_conversion in KokoroEngine, which uses a dual-loop (Generation & Playback) architecture. - Buffer Management: The engine maintains a buffer of generated audio ahead of the playback point (targeted at ~2 minutes) to ensure smooth streaming. - Interactive Playback: Uses winsound in a dedicated thread to play audio chunks as they are generated. Playback can be immediately interrupted using the "Cancel" button. - Save State on Stop: If generation is cancelled or stopped, the engine autonomously: - Combines all played and generated audio into a single .wav file. - Exports all unplayed text (including the segment currently being played when stopped) to a _remaining.txt file, allowing users to resume later. - Format Compatibility: Automatically ensures JIT chunks use WAV format for compatibility with the Windows playback system, regardless of the global output setting. - Lexicon & FX Support: JIT mode fully supports the existing Lexicon (pronunciation dictionary) and Audio FX pipeline.

CoffeeMethod7 months agoe95b6b1

Added more mixing operations to the Custom Voice feature.

CoffeeMethod7 months ago5b8cb44

Implement caching system for repeated generations

CoffeeMethod7 months agoc8c0541

Added Audio FX.

CoffeeMethod8 months agob185b6c