JOIN THE GLOBAL VOICE AI GATHERING 👉
    HERE
    Voice AI / Space
    Submit
    Back
    Soniqo

    Soniqo

    Visit Website
    Vertical
    Tech
    Open Source
    On device
    Diarisation
    Voice Cloning
    Productivity
    Note Taker

    Open-source, fully offline on-device speech AI for transcription and cloning.

    Soniqo banner

    About Soniqo

    Soniqo: On-Device Speech AI

    Soniqo is an open-source, fully offline on-device speech AI platform designed for real products. It provides diarized transcription, zero-shot voice cloning, and long-form speech synthesis without relying on cloud APIs. Operating under an Apache 2.0 license, Soniqo ensures no data leaves the device and eliminates per-minute pricing, running efficiently on Apple Silicon, Android, Windows, and embedded Linux.

    Key Features

    • Fully Offline Processing: Operates entirely on-device with no cloud APIs, ensuring data privacy and zero per-minute pricing.
    • Cross-Platform Compatibility: Runs on Apple Silicon, Android, Windows, and embedded Linux, with memory budgets as low as 1.2 GB on mobile devices.
    • Comprehensive Model Stack: Includes over thirty models for speech-to-text, text-to-speech, audio analysis, and LLM integration, featuring models like Whisper, Parakeet, CosyVoice 3, and Silero VAD.
    • Real-Time Performance: Delivers low-latency processing for streaming transcription, voice activity detection, and full-duplex speech-to-speech interactions.

    Use Cases

    • Conversational Voice Agents: Build voice-first interfaces, including full-duplex speech-to-speech and wake-word-driven compositional pipelines.
    • Audio Understanding: Turn audio into structured text with real-time streaming for live captions, batch processing for archives, and speaker diarization.
    • Content Creation: Synthesize speech to clone voices in seconds, narrate audiobooks, or cast multi-speaker podcasts fully offline.

    Getting Started

    Website: https://soniqo.audio/

    Installation is available via Homebrew for Apple devices and Gradle for Android.

    Soniqo equips developers with a robust, open-source SDK to build secure, offline voice applications across multiple platforms using state-of-the-art speech models.