selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

Alternatives to mithi-v1-flash

120 open-weight models that do the same job as mithi-v1-flash (99M TTS), ranked by how close they are in task and size, then by popularity and how recent they are. Licence and VRAM are shown for each.

the original
Oomni logo

mithi-v1-flash

Oomni·TTS

≈0.7 GB

A small text-to-speech model for 11 Indian languages, with two voices: Satvik (male) and Charvi (female).

99MApache 2.0
1.4KNov 2026
--vram
any
--gpus
--quant
  1. fishaudio logo

    s1-mini

    fishaudio·TTS

    ? GB

    FishAudio S1 is a leading text-to-speech (TTS) model trained on more than 2 million hours of audio data in multiple languages.

    ?BCC BY-NC-SA 4.0
    2.2KMay 2025gguf
    • Same category
    vs
  2. BreezeBlue logo

    Breeze-TTS-2

    BreezeBlue·TTS

    ≈4.6 GB

    Breeze TTS 2 is an open-weight text-to-speech model built for real-time interaction. It ranks #1 among open-weight models on the Artificial Analysis TTS leaderboard, while outperforming frontier proprietary systems.

    3.5Bbreezeblue-research-and-non-commercial-license
    26.9KAug 2026gguf
    • Same category
    • Larger (3.5B vs 99M)
    vs
  3. ByteDance logo

    MegaTTS3

    ByteDance·TTS

    ? GB

    \[!TIP] \[IMPORTANT] For security issues, we do not upload the parameters of WaveVAE encoder to the above links. You can only use the pre-extracted latents from link1 for inference. If you want to synthesize speech for speaker A, you need "A.wav" and "A.npy" in the same directory.

    ?BApache 2.0
    2.7KMar 2025
    • Same category
    • Commercial use OK
    vs
  4. m-a-p logo

    YuE2-3B

    m-a-p·TTS

    ≈4.8 GB

    YuE2 is an open music generation model with frontier song quality competitive with Suno v5/v6. Turn lyrics and a style prompt into a complete song with vocals and accompaniment, then shape its melody and chords through an editable score.

    3.6B24K ctx CC BY-NC 4.0
    20.3KSept 2026gguf
    • Same category
    • Larger (3.6B vs 99M)
    vs
  5. inclusionAI logo

    Ming-3.0-tts-0.5B

    inclusionAI·TTS

    ≈3.8 GB

    This repository contains Ming-3.0-tts-0.5B, the dense Ming-3.0-tts checkpoint built on Qwen2.5-0.5B. The checkpoint includes the backbone, DiT head, audio tokenizer, speaker encoder and text tokenizer.

    1.5BMIT
    18.6KSept 2026gguf
    • Same category
    • Larger (1.5B vs 99M)
    • Commercial use OK
    vs
  6. fishaudio logo

    fish-speech-1.5

    fishaudio·TTS

    ? GB

    Fish Speech V1.5 is a leading text-to-speech (TTS) model trained on more than 1 million hours of audio data in multiple languages.

    ?BCC BY-NC-SA 4.0
    5.2KNov 2024
    • Same category
    vs
  7. SparkAudio logo

    Spark-TTS-0.5B

    SparkAudio·TTS

    ≈1.6 GB

    Spark-TTS is an advanced text-to-speech system that uses the power of large language models (LLM) for highly accurate and natural-sounding voice synthesis. It is designed to be efficient, flexible, and powerful for both research and production use.

    500MCC BY-NC-SA 4.0
    2.8KFeb 2025gguf
    • Same category
    • Larger (500M vs 99M)
    vs
  8. mehdi-hf logo

    cilla-mini

    mehdi-hf·TTS

    ≈0.6 GB

    Gated, non-commercial (CC BY-NC 4.0): ask for access on this page. Cilla Mini is the 18M model for CPUs, phones and browsers: Cilla shrunk with Kyutai's small-model recipe (8 layers, a 64-dim interface, a one-step drifting sampler head, the 2.6M Mimi decoder), guidance baked in.

    28MCC BY-NC 4.0
    0Nov 2026
    • Same category
    • Smaller (28M vs 99M)
    • Newer release
    vs
  9. microsoft logo≈2.8 GB

    VibeVoice-Realtime is a lightweight real‑time text-to-speech model supporting streaming text input and robust long-form speech generation.

    1BMIT
    129.4KDec 2025gguf
    • Same category
    • Larger (1B vs 99M)
    • Commercial use OK
    vs
  10. cloud0day3 logo

    antalia-mini

    cloud0day3·TTS

    ≈0.5 GB

    Türkçe özet. Antalia-2 Mini, tek sesli, küçük bir Türkçe metinden konuşmaya (TTS) modelidir: toplam 7,6 milyon parametre, 48 kHz çıkış, sıradan bir işlemcide gerçek zamandan çok daha hızlı.

    8MApache 2.0
    12.2KOct 2026
    • Same category
    • Smaller (8M vs 99M)
    • Commercial use OK
    vs
  11. FunAudioLLM logo

    Fun-CineForge

    FunAudioLLM·TTS

    ? GB

    🎬 Fun-CineForge: A Unified Dataset Pipeline and Model for Zero-Shot Movie Dubbing in Diverse Cinematic Scenes

    ?BApache 2.0
    109Mar 2026
    • Same category
    • Commercial use OK
    vs
  12. n4ze3m logo

    tiny-voca

    n4ze3m·TTS

    ≈0.5 GB

    tiny-voca is a small English text-to-speech model with two voices, F2 (female) and M2 (male). The complete model, including the waveform generator, has 4.5M parameters and was trained from scratch on about 18 hours of synthetic speech. It runs about 70x faster than real time on four CPU threads.

    5MApache 2.0
    7.9KNov 2026
    • Same category
    • Smaller (5M vs 99M)
    • Commercial use OK
    • Newer release
    vs
  13. microsoft logo

    VibeVoice-1.5B

    microsoft·TTS

    ≈6.5 GB

    VibeVoice is a novel framework designed for generating expressive, long-form, multi-speaker conversational audio, such as podcasts, from text. It addresses significant challenges in traditional Text-to-Speech (TTS) systems, particularly in scalability, speaker consistency, and natural turn-taking.

    2.7BMIT
    214.3KAug 2025gguf
    • Same category
    • Larger (2.7B vs 99M)
    • Commercial use OK
    vs
  14. myshell-ai logo

    MeloTTS-Chinese

    myshell-ai·TTS

    ? GB

    MeloTTS is a high-quality multi-lingual text-to-speech library by MyShell.ai. Supported languages include:

    ?BMIT
    25.1KFeb 2024
    • Same category
    • Commercial use OK
    vs
  15. fishaudio logo

    s2-pro

    fishaudio·TTS

    ≈5.9 GB

    Fish Audio S2 Pro is a leading text-to-speech (TTS) model with fine-grained inline control of prosody and emotion. Trained on over 10M+ hours of audio data across 80+ languages, the system combines reinforcement learning alignment with a dual-autoregressive architecture.

    4.6Bfish-audio-research-license
    43KMar 2026gguf
    • Same category
    • Larger (4.6B vs 99M)
    vs
  16. inclusionAI logo

    Ming-3.0-tts-tiny

    inclusionAI·TTS

    ≈11 GB

    This repository contains Ming-3.0-tts-tiny, the Ming-3.0-tts mixture-of-experts checkpoint built on Ling V3. Its Ling V3 backbone has 7.9B total parameters and activates about 1.3B parameters per token.

    9BA2.5BMIT
    9.4KSept 2026gguf
    • Same category
    • Larger (9B vs 99M)
    • Commercial use OK
    vs
  17. ConwayResearch logo

    Underdog-Woof

    ConwayResearch·TTS

    ? GB

    Underdog Woof is a small English text-to-speech model with two voices: Juno and Mira. It reads phone numbers, dates, prices and codes the way people say them, and it runs on a laptop CPU with nothing but numpy, from a single 40.6 MB file. Try it in your browser at underdog.ai/woof.

    ?Bunderdog-non-commercial-1.0
    10Nov 2026
    • Same category
    vs
  18. 2Noise logo

    ChatTTS

    2Noise·TTS

    ? GB

    We are also training larger-scale models and need computational power and data support. If you can provide assistance, please contact OPEN-SOURCE@2NOISE.COM. Thank you very much.

    ?BCC BY-NC 4.0
    8.2KMay 2024
    • Same category
    vs
  19. myshell-ai logo

    MeloTTS-Spanish

    myshell-ai·TTS

    ? GB

    MeloTTS is a high-quality multi-lingual text-to-speech library by MyShell.ai. Supported languages include:

    ?BMIT
    13.3KFeb 2024
    • Same category
    • Commercial use OK
    vs
  20. kyutai logo

    tts-1.6b-en_fr

    kyutai·TTS

    ≈4.1 GB

    See also the pre-print research paper, the project page, the Colab example, the GitHub repository, and the repository of voices.

    1.6BCC BY 4.0
    133.5KJun 2025gguf
    • Same category
    • Larger (1.6B vs 99M)
    • Commercial use OK
    vs
  21. openbmb logo

    VoxCPM1.5

    openbmb·TTS

    ≈2.3 GB

    Key Improvements: - 🔊 Higher Quality: 44.1kHz sampling rate preserves more high-frequency details for better voice cloning - ⚡ More Efficient: Reduced token rate (6.25Hz) lowers computational cost while maintaining performance - 🎓 Fine-tuning Support: Train personalized voice models with SFT or…

    802MApache 2.0
    6KDec 2025gguf
    • Same category
    • Larger (802M vs 99M)
    • Commercial use OK
    vs
  22. OuteAI logo

    OuteTTS-0.1-350M

    OuteAI·TTS

    ≈1.3 GB

    OuteTTS-0.1-350M is a novel text-to-speech synthesis model that leverages pure language modeling without external adapters or complex architectures, built upon the LLaMa architecture using our Oute3-350M-DEV base model, it demonstrates that high-quality speech synthesis is achievable through a…

    362M4K ctx CC BY 4.0
    95Nov 2024
    • Same category
    • Larger (362M vs 99M)
    • Commercial use OK
    vs
  23. inclusionAI logo

    Ming-3.0-audio-tiny

    inclusionAI·TTS

    ≈11 GB

    Ming-3.0-audio-tiny is a compact native audio-language model from inclusionAI. It accepts audio together with text instructions and supports speech recognition, spoken question answering, audio understanding, audio reasoning, and optional synthesized-speech responses.

    8.9BA2.1BMIT
    5.1KSept 2026gguf
    • Same category
    • Larger (8.9B vs 99M)
    • Commercial use OK
    vs
  24. sbintuitions logo

    sarashina2.2-tts

    sbintuitions·TTS

    ≈2.3 GB

    Sarashina2.2-TTS is a Japanese-centric text-to-speech system built on a large language model, developed by SB Intuitions. It supports Japanese and English, delivering high pronunciation accuracy, naturalness, and stability across diverse speaking styles, with zero-shot voice generation support.

    810M8K ctx Unknown licence
    2.2KApr 2026gguf
    • Same category
    • Larger (810M vs 99M)
    vs