/* TEAM */

  Developer: Matthieu Veinhard
  Contact: support [at] thoth-app.com
  Location: France

  Thoth is built by one person. No funding, no team, no bot joins your call.

/* THANKS: ON-DEVICE AI */

  Thoth runs entirely on your device. It stands on open models and
  open-source work from many people. Thank you.

  On-device summary models (local LLMs):
    Gemma 3 (Google DeepMind)
    Qwen (Alibaba Qwen team)
    Ministral / Mistral (Mistral AI)
    Phi 3.5 (Microsoft)

  Speech and speaker models:
    Whisper (OpenAI), run on device via WhisperKit (Argmax)
    Parakeet TDT and Parakeet EOU (NVIDIA), via FluidAudio
    pyannote speaker diarization (pyannote.audio)

  Frameworks and libraries:
    MLX (Apple)
    llama.cpp / ggml (Georgi Gerganov and contributors)
    swift-transformers and swift-jinja (Hugging Face)

/* SITE */

  Last update: 2026/06/16
  Languages: English, French
  Standards: HTML5, CSS3
  Components: hand-written HTML and CSS, vanilla JavaScript, self-hosted fonts
  Hosting: Cloudflare Pages
  Software: written by hand in a text editor