BryceWG/BiBi-Keyboard

说点啥（BiBi Keyboard）:一个基于 Kotlin 的 Android 平台的 LLM 与 ASR 语音输入法键盘应用 An LLM ASR voice input method keyboard application for the Android platform based on Kotlin

What it solves

BiBi Keyboard is an AI-powered voice input method for Android that aims to make voice typing more natural and efficient. It solves the common frustrations of traditional voice input by providing high-accuracy recognition, flexible engine options, and AI-driven text correction.

How it works

The app functions as a system-wide Android keyboard (IME) or a floating ball overlay. It captures audio via a long-press recording mechanism and processes it using one of 18 supported ASR (Automatic Speech Recognition) providers, including 12 cloud-based services and 6 local offline models (such as SenseVoice and FunASR Nano). To improve quality, it uses LLMs for post-processing to correct recognition errors and allows users to edit text via voice commands in a dedicated AI editing panel.

Who it’s for

Android users who prefer voice input over typing, users requiring offline voice recognition capabilities, and power users who want to integrate AI voice input into their existing third-party keyboards via LSPosed/LSPatch.

Highlights

  • Hybrid ASR Support: Supports both cloud and local offline models (e.g., Qwen3-ASR, FireRedASR) for reliability and privacy.
  • Floating Ball Input: Allows voice input to be used across any installed keyboard by automatically inserting the recognized text.
  • AI Post-Processing: Uses LLMs to refine and correct the transcribed text after recognition.
  • Extensive Integration: Features an IME Bridge for third-party keyboard linkage and supports external voice input interfaces (AIDL).
  • Modern UI: Built with Jetpack Compose and Material 3 / Miuix styles.

Related

  • Project
    xifan2333/fcitx5-vinput

    A voice input addon for Fcitx5 that enables local or cloud-based speech-to-text with LLM-powered post-processing and text modification.

  • Project
    CSCB/vibe-mouse

    A customizable interaction layer that maps AI IDE actions to various hardware devices and voice commands to streamline AI-assisted coding workflows.

  • Project
    woheller69/whisperIME

    An offline voice recognition input method for Android based on the Whisper engine, providing privacy-focused speech-to-text and translation.

  • Project
    ximeiorg/Xime

    An Android input method based on the Rime engine that provides efficient Chinese text input via Wubi and Pinyin, enhanced by on-device AI for predictive text and offline speech recognition.

  • Project
    LeanBitLab/LeanType

    A private, open-source Android keyboard that integrates cloud and on-device AI for proofreading, translation, Whisper voice typing, and OCR text extraction.