| DramaBox - Expressive TTS (Kaggle) |
Emotional dialogue, laughter, whispers, and instant voice cloning powered by Resemble AI DramaBox (LTX audio DiT + Gemma-3 12B) with dual GPU parallel denoising. Optimized for Kaggle GPU T4 x2. |
 |
 |
| MiniMax H3 Turbo - INT8 Video & Audio Generator (Kaggle) |
20B foundation model generating synchronized video & 32 kHz stereo audio from text, images, or reference inputs via 4-step PDMD/DMAD distillation & INT8 tensor-parallel DiT. Optimized for Kaggle GPU T4 x2. |
|
|
| Fish Audio S2.1 Pro Studio - Expressive Voice Cloning & TTS (Colab) |
Dual-autoregressive speech synthesis across 83+ languages with instant zero-shot voice cloning and paralinguistic emotion control. 100% free cloud API & CPU-friendly (no GPU required). |
 |
|
| Qwen-Image 2.1 Uncensored - Create, Edit & Multi-Reference (Colab) |
High-speed text-to-image, instruction editing, multi-reference composition (up to 10 images), inpainting, outpainting & transparent PNGs via INT8 Tensor Cores. Optimized for Colab Free T4 GPU. |
|
 |
| AuK v1.0 - Unified Audio Generation & Editing (Kaggle) |
1.5B foundational base audio model for high-quality text-to-audio, text-to-music & audio editing. Optimized for Kaggle GPU T4 x2. |
|
|
| YuE2-3B - Frontier Full-Song Music Generation (Kaggle) |
High-fidelity 48 kHz stereo music generation with unified symbolic planning (ABC) & flow matching. Optimized for Kaggle GPU T4 x2. |
|
|
| LTX-2.5 22B Distilled - Dual GPU Video Generator (Kaggle) |
Official LTX-2.5 22B Distilled with 8-step high-speed video generation & synchronized stereo audio. Optimized for Kaggle GPU T4 x2. |
|
|
| HYPIR - 4K Image Upscaler & Restoration (Colab) |
Single-step high-fidelity image restoration and 4K upscaling using diffusion-yielded score priors. Optimized for Colab Free T4 GPU. |
|
|
| MOSS-TTS v1.5 - Standalone Foundation Model (Kaggle) |
High-fidelity 48 kHz stereo TTS and zero-shot voice cloning. Optimized for Kaggle T4 x2 GPU. |
|
|
| ScenA Audio - Expressive Speech Generator (Colab) |
✅ Fixed & working as of September 2026. Zero-Shot Voice Cloning, Intent-Aware TTS, and Multi-Speaker Dialogue using the ScenA Audio model. Optimized for Colab Free T4 GPU. |
|
|
| Krea 2 Turbo - Fast Text-to-Image Generator (Kaggle) |
✅ Fixed & working as of September 2026. Ultra-fast text-to-image generation powered by the Krea 2 Turbo model. Optimized for Kaggle GPU T4 x2. |
|
|
| LTX-2.3 22B MSR Ref Distilled 1.1 (Kaggle) |
✅ Fixed & working as of September 2026. Multi-Reference Consistent Video Generation (Ingredients-to-Video) utilizing LTX-2.3 22B Distilled 1.1 + LiconStudio MSR LoRA. Optimized for Kaggle GPU T4 x2. |
|
|
| Dots.TTS - Zero-Shot Autoregressive TTS (Colab) |
✅ Fixed & working as of September 2026. 2B Parameter continuous autoregressive TTS with SOAR alignment & MeanFlow distillation for fast voice cloning. Optimized for Colab Free T4 GPU. |
|
|
| SeedVC - Voice Changer & Singing Cloner (Colab) |
✅ Checked & working as of September 2026. Zero-Shot Voice Conversion · Vocal Demixing · Song Cover Cloning. Powered by SeedVC & RoFormer. Optimized for Colab Free T4 GPU. |
|
|
| Scenema Audio Expressive Speech Generator (Kaggle) |
✅ Fixed & working as of September 2026. Zero-Shot Voice Cloning, Intent-Aware TTS, and Multi-Speaker Dialogue using the Scenema Audio model. Optimized for Kaggle GPU T4 x2. |
|
|
| LTX-2.3 22B Distilled 1.1 Q4 Text & Image-to-Video (Kaggle) |
✅ Fixed & working as of September 2026. High-quality text-to-video & image-to-video generation with synchronized audio using LTX-2.3 22B Distilled 1.1 Q4. Optimized for Kaggle GPU T4 x2. |
|
|
| LTX-2.3 22B Audio-to-Video Q4 (Kaggle) |
✅ Fixed & working as of September 2026. Audio-driven video generation (talking head / lip sync) powered by LTX-2.3 Q4. Optimized for Kaggle GPU T4 x2. |
|
|
| ACE-Step 1.5 XL Turbo 4B (Kaggle) |
✅ Checked & working as of September 2026. High-quality AI music generation with the 4B XL Turbo model. Optimized for Kaggle GPU T4 x2. |
|
|
| VoxCPM2 - Multilingual TTS (Colab) |
✅ Checked & working as of September 2026. 2B Parameters · 30 Languages · 48kHz Output · Voice Design & Cloning. Optimized for Colab T4 GPU. |
|
|
| OmniVoice - 600+ Language Zero-Shot TTS (Colab) |
✅ Checked & working as of September 2026. Voice Cloning · Voice Design · Auto Voice · 600+ Languages. Powered by k2-fsa/OmniVoice. Optimized for Colab Free T4. |
|
|
| Cohere Transcribe (Colab) |
✅ Fixed & working as of September 2026. Now powered by ungated model weights (no HF token required). State-of-the-art 2B Conformer ASR for 14 languages & long-form audio. Optimized for Colab Free T4. |
|
|
| LTX-2.3 22B Distilled (Kaggle) |
✅ Fixed & working as of September 2026. Previously configured for P100, now updated for Kaggle GPU T4 x2 with Wan2GP engine & INT8 kernels for faster output generation. |
|
|
| SoulX FlashHead - Talking Head Generator (Colab) |
✅ Fixed & working as of September 2026. Audio-driven AI talking head generator featuring fast Lite & high-quality Pro models. Optimized for Colab Free T4 GPU. |
|
|
| ACE-Step 1.5 - Music Generator & LoRA Studio (Colab) |
✅ Fixed & working as of September 2026. Custom Gradio music studio with full LoRA training, text-to-music, remixing & audio repainting. Optimized for Colab Free T4 GPU. |
|
|
| MOSS-TTS 1.7B (Colab) |
✅ Fixed & working as of September 2026. Zero-shot voice cloning & high-fidelity TTS. Optimized for Colab Free T4 GPU. |
|
|
| HeartMuLa 3B - Standalone Music Generator (Colab) |
✅ Fixed & working as of September 2026. Free & open-source AI music generation. Optimized for Colab Free T4 GPU. |
|
|
| Qwen3-TTS 1.7B (Colab) |
✅ Fixed & working as of September 2026. Advanced Text-to-Speech AI featuring voice cloning, custom voice & voice design. Optimized for Colab Free T4 GPU. |
|
|
一个精心整理的免费即开即用 Google Colab 和 Kaggle 笔记本合集，每个笔记本封装一个不同的 AI 工具，涵盖文本转语音、声音克隆、图像生成、视频生成或音乐生成。