MikeTrendsTrends right now
AI: reposAPI

search

text-to-speech

Trends

  1. 1
    Fish Audio Launches Drama 3, Top Open-Weights TTS Model●𝕏xSE2.9K5 d ago

    Fish Audio has released Drama 3, a text-to-speech model the company says now leads the open-weights field, meaning anyone can download and run the model weights themselves. The launch is drawing attention from developers and AI enthusiasts comparing it to closed competitors, with discussions centring on voice quality, licensing terms, and what it means for free and commercial speech-generation tools.

  2. 2
    Babytalk brings offline speech recognition to ESP32 microcontrollersβ–ΌShow HN: Babytalk: Offline speech to text and text to speech on ESP32YhnWarTerrorism142 d ago

    A developer has released Babytalk, an open-source project that runs speech-to-text and text-to-speech entirely offline on an ESP32 microcontroller. The project, shared on GitHub, lets hobbyists build voice-controlled devices without cloud services or internet connectivity. Early reactions highlight how impressive it is to fit speech processing on such low-cost, low-power hardware.

  3. 3
    New Tool Turns OpenAI Subscription Into Audiobook Maker●Show HN: Use your OpenAI subscription to create audiobooksYhnCultureMusic1595 d ago

    A developer has launched a service that lets people use their existing OpenAI subscription to convert text, such as PDFs, into audiobooks with AI-generated narration. The tool, shared with the Hacker News community, has drawn attention for making audiobook creation accessible without extra text-to-speech costs, with early reactions focusing on its practicality and voice quality.

  4. 4
    German Philosophy Project Launches English Audio Editionβ–ΌπŸŽ§ Don’t feel like reading? Starting now, # philosophies_de is also available directly for your # Englishspeaking ears!πŸ‘‚MmastodonTechnologyAI24 d ago

    Philosophies_de, a German-language philosophy and science project, has begun publishing an English-language audio version of its content on Spotify. The audio edition is produced using text-to-speech technology, with artificial intelligence converting the written material into spoken episodes for listeners who prefer audio over reading. The announcement is circulating among philosophy, science and podcast audiences online.

  5. 5
    VoxCPM2 brings voice cloning to home PCsβ—πŸ’‘ VoxCPM2 porta sul PC di casa la clonazione della voce e la creazione di voci da una semplice descrizione, con trenta lMmastodonTechnologySoftware36 d ago

    VoxCPM2, an open-source text-to-speech system, allows users to clone voices and generate new ones from simple text descriptions directly on a home computer. It supports thirty languages, including Italian, and produces professional-quality audio. Tech communities are sharing it as an accessible alternative to commercial voice synthesis services.

  6. 6
    Kokoro 82M trails ElevenLabs in open-source speech rankingβ—πŸŽ¨ Where open-source stands in text to speech: #54 Kokoro 82M v1.0 β€” the best model you can actually download 263 ELO poiMmastodonTechnologySoftware26 d ago

    A text-to-speech leaderboard entry ranks Kokoro 82M v1.0, described as the strongest open-source model available for download, at number 54 β€” 263 ELO points behind ElevenLabs' Eleven v4 Turbo. The ranking is based on blind human preference tests rather than vendor benchmarks, highlighting the gap between downloadable open models and leading commercial systems.

  7. 7
    Hobbyist runs voice cloning experiments on decade-old PC●I just wanted a local TTS tool to go along with Whisper. Somehow, I ended up experimenting with voice cloning on my almoMmastodonTechnologySoftware16 d ago

    A user says they set out to find a local text-to-speech tool to pair with Whisper but ended up testing voice cloning on their nearly 14-year-old PC. They compiled audio.cpp with the Vulkan backend from source and tried the Audio8 TTS Preview 0.6B Q8_0 model, noting odd results. The anecdote highlights how accessible consumer AI speech tools have become on modest hardware.

  8. 8
    Concat offers free offline video editor without watermarks●Cerchi un editor video gratuito, senza watermark e capace di funzionare offline? Concat punta a offrire strumenti per moMmastodonTechnologySoftware16 d ago

    Concat is being promoted as a free, open-source video editor that works offline, adds no watermarks, and keeps all footage on the user's device. It combines editing tools with automatic subtitles, text-to-speech and video effects, and is aimed particularly at Linux users looking for an alternative to subscription-based or cloud-dependent editing software.

Repos