AI Search's banner
AI Search's profile picture

AI Search

@aisearchio • 27,190 subscribers

Sharing AI stuff 🚀 760K subs Nature lover, PhD, investor, shitposter https://t.co/0Ms1yEyG5f

Shorts

Seedance 2.0. Absolute madness

Seedance 2.0. Absolute madness

108,053 views

Here's a fast preview decoder for MiniMax-H3. It converts the model’s hidden image data into a rough preview while the video is generating, so you can quickly see the progress without waiting for the final result.

Here's a fast preview decoder for MiniMax-H3. It converts the model’s hidden image data into a rough preview while the video is generating, so you can quickly see the progress without waiting for the final result.

32,201 views

Alibaba doesn't stop cooking. They just dropped a REALTIME, infinite length video generator. Based on Wan, 20 fps, with dialogue

Alibaba doesn't stop cooking. They just dropped a REALTIME, infinite length video generator. Based on Wan, 20 fps, with dialogue

78,776 views

No need to wait for the next season. Just make it in Seedance 2.0

No need to wait for the next season. Just make it in Seedance 2.0

61,313 views

Y'all aint ready for Seedance 2.0. Holy shit.

Y'all aint ready for Seedance 2.0. Holy shit.

58,699 views

This AI allows you to stream as anyone in realtime. Free & open source, works with regular GPU. It was a pain to install but I eventually got it to work. Full tutorial:

This AI allows you to stream as anyone in realtime. Free & open source, works with regular GPU. It was a pain to install but I eventually got it to work. Full tutorial:

59,364 views

Seedance 2.0 is too much fun. What prompts would you like to test? Comment below and I'll select some to generate 👇

Seedance 2.0 is too much fun. What prompts would you like to test? Comment below and I'll select some to generate 👇

31,208 views

Krea 2 is a very impressive and uncensored model. Full tutorial & review:

Krea 2 is a very impressive and uncensored model. Full tutorial & review:

15,562 views

One of them is MiniMax H3 by MiniMax Design (H3) Here's a preliminary test.

One of them is MiniMax H3 by MiniMax Design (H3) Here's a preliminary test.

11,810 views

Minimax H3 is pretty good at making MVs. Because it's multimodal, it can take in audio, images, and video. Here, I uploaded a song made from ACE Step and images from Nano Banana. It's able to sync the video to the audio quite well. Prompt:

Minimax H3 is pretty good at making MVs. Because it's multimodal, it can take in audio, images, and video. Here, I uploaded a song made from ACE Step and images from Nano Banana. It's able to sync the video to the audio quite well. Prompt:

11,476 views

NVIDIA's PiD is a new pixel diffusion decoder for high-res image models. It skips decode-then-upscale stop, making sharper outputs faster. > Directly generates 4K images > Up to 5.9x faster than SeedVR2 > Free & open weights

NVIDIA's PiD is a new pixel diffusion decoder for high-res image models. It skips decode-then-upscale stop, making sharper outputs faster. > Directly generates 4K images > Up to 5.9x faster than SeedVR2 > Free & open weights

15,083 views

Finally! Long continuous AI video is here Free & open source. Infinite length Prompts: she looks at the camera and smiles she drinks the glass of water she stands up and walks away Made with SVI 2.0 Pro Full tutorial:

Finally! Long continuous AI video is here Free & open source. Infinite length Prompts: she looks at the camera and smiles she drinks the glass of water she stands up and walks away Made with SVI 2.0 Pro Full tutorial:

24,030 views

LongCat-Video-Avatar 1.5 is a major upgrade for AI avatar generation. > Better lip-sync + mouth-shape accuracy > Stable identities across long videos & multi-person scenes > Smooth full-body motion, singing, animation, & hand interactions > Outperforms HeyGen, Kling Avatar 2.0, & OmniHuman-1.5 in consistency tests Free & open source

LongCat-Video-Avatar 1.5 is a major upgrade for AI avatar generation. > Better lip-sync + mouth-shape accuracy > Stable identities across long videos & multi-person scenes > Smooth full-body motion, singing, animation, & hand interactions > Outperforms HeyGen, Kling Avatar 2.0, & OmniHuman-1.5 in consistency tests Free & open source

12,645 views

Videos

LIVE
1.2k

Anya Rossi

sweetdream.ai

SweetDream.aiSponsored•Livecam
Streaming Now

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

HD live stream
Exclusive private shows
1.2k viewers online

Free preview available • Premium content

New open source music generator, YuE2 Even beats Mureka v9 and Minimax 3 on various benchmarks
3:09
aisearchio's profile picture

New open source music generator, YuE2 Even beats Mureka v9 and Minimax 3 on various benchmarks

AI Search

430,322 views • 1 month ago

Drama 3 by Fish Audio is now the top open TTS model. Allows meta tags within your transcript to control emotion, direction, SFX. Pretty good at voice cloning too.
0:49
aisearchio's profile picture

Drama 3 by Fish Audio is now the top open TTS model. Allows meta tags within your transcript to control emotion, direction, SFX. Pretty good at voice cloning too.

AI Search

21,510 views • 3 days ago

Ideogram 4.5 edits even better than GPT Image 2.5. And it will be open weights! Can't wait to run this locally
1:12
aisearchio's profile picture

Ideogram 4.5 edits even better than GPT Image 2.5. And it will be open weights! Can't wait to run this locally

AI Search

51,624 views • 18 days ago

GPT 6 Astra crushes 3D game design. Here's a sneak peek of one of my tests.
0:52
aisearchio's profile picture

GPT 6 Astra crushes 3D game design. Here's a sneak peek of one of my tests.

AI Search

43,242 views • 1 month ago

NVIDIA's LocateAnything is a new vision model for grounding and detection. Very performant and accurate! > 10x faster than Qwen3-VL > 138M queries + 785M boxes > GUI, OCR, docs, dense detection > Free & open source
1:20
aisearchio's profile picture

NVIDIA's LocateAnything is a new vision model for grounding and detection. Very performant and accurate! > 10x faster than Qwen3-VL > 138M queries + 785M boxes > GUI, OCR, docs, dense detection > Free & open source

AI Search

122,236 views • 4 months ago

One of my favorite abilities of Seedance 2.0 You can use 3D motion + character references and get an insane generation like this. More use cases:
0:26
aisearchio's profile picture

One of my favorite abilities of Seedance 2.0 You can use 3D motion + character references and get an insane generation like this. More use cases:

AI Search

151,873 views • 8 months ago

Kimi K3 is a total freak. It made a full V8 engine using Blender MCP in just one prompt. Plus some other fun stuff. Full review
0:29
aisearchio's profile picture

Kimi K3 is a total freak. It made a full V8 engine using Blender MCP in just one prompt. Plus some other fun stuff. Full review

⚡AI Search⚡

59,447 views • 3 months ago

Another cool project with Minimax H3. H3-World turns Minimax into an interactive world generator.
0:31
aisearchio's profile picture

Another cool project with Minimax H3. H3-World turns Minimax into an interactive world generator.

AI Search

28,164 views • 1 month ago

The quality of this new speedup is great! Here are some comparisons with Minimax base and FastH3. Waiting for Lord Kijai to add ComfyUI support
0:43
aisearchio's profile picture

The quality of this new speedup is great! Here are some comparisons with Minimax base and FastH3. Waiting for Lord Kijai to add ComfyUI support

AI Search

22,167 views • 1 month ago

LTX 2.5 vs Minimax H3
0:20
aisearchio's profile picture

LTX 2.5 vs Minimax H3

AI Search

29,584 views • 2 months ago

SCAIL-2 is a new SOTA AI that transfers motion from one video onto another. Works with multiple characters, no need for pose skeletons or masks. Free & open-weights. Interestingly, the author affiliations include Z AI, the team behind GLM.
0:31
aisearchio's profile picture

SCAIL-2 is a new SOTA AI that transfers motion from one video onto another. Works with multiple characters, no need for pose skeletons or masks. Free & open-weights. Interestingly, the author affiliations include Z AI, the team behind GLM.

AI Search

43,567 views • 4 months ago

Open-source SUNO is here! Introducing HeartMula: free & offline AI music generator Full tutorial:
1:15
aisearchio's profile picture

Open-source SUNO is here! Introducing HeartMula: free & offline AI music generator Full tutorial:

AI Search

77,777 views • 9 months ago

Raon-OpenTTS-1B is a new open text-to-speech model for zero-shot voice cloning. It has best overall similarity score and lowest error rate vs other leading TTS models.
0:42
aisearchio's profile picture

Raon-OpenTTS-1B is a new open text-to-speech model for zero-shot voice cloning. It has best overall similarity score and lowest error rate vs other leading TTS models.

AI Search

22,657 views • 2 months ago

Bytedance drops an open-source Gemini Omni!!! Bernini is a new AI video generation + editing framework. > Edit videos with text prompts > Image/video references > Code available
0:31
aisearchio's profile picture

Bytedance drops an open-source Gemini Omni!!! Bernini is a new AI video generation + editing framework. > Edit videos with text prompts > Image/video references > Code available

AI Search

44,413 views • 4 months ago

More YuE2 demos. From some initial tests, it does sound cleaner and more versatile than Minimax 3 and Acestep.
8:00
aisearchio's profile picture

More YuE2 demos. From some initial tests, it does sound cleaner and more versatile than Minimax 3 and Acestep.

AI Search

14,955 views • 1 month ago

GPT 6 Astra composing an EDM track from scratch in my DAW
3:00
aisearchio's profile picture

GPT 6 Astra composing an EDM track from scratch in my DAW

AI Search

16,378 views • 1 month ago

SolarWM turns Minimax H3 into realtime interactive worlds. It can keep generating for over an hour.
0:48
aisearchio's profile picture

SolarWM turns Minimax H3 into realtime interactive worlds. It can keep generating for over an hour.

AI Search

16,721 views • 1 month ago

GLM 5.2 by Z.ai is even bigger than the DeepSeek moment. In the past few months, open source AI had been quite stagnant. New releases were underwhelming and still far behind the top closed models. But GLM 5.2 is different. It gets surprisingly close to GPT and Opus-level quality, while being much cheaper and open-weight. It just gets things done with very minimal hand holding. Full review
0:43
aisearchio's profile picture

GLM 5.2 by Z.ai is even bigger than the DeepSeek moment. In the past few months, open source AI had been quite stagnant. New releases were underwhelming and still far behind the top closed models. But GLM 5.2 is different. It gets surprisingly close to GPT and Opus-level quality, while being much cheaper and open-weight. It just gets things done with very minimal hand holding. Full review

AI Search

36,507 views • 4 months ago

Controlnet for Minimax H3 is here!
0:33
aisearchio's profile picture

Controlnet for Minimax H3 is here!

AI Search

19,041 views • 1 month ago

Krea 2 full review & installation tutorial (I'll be uploading more full videos on X. Please support me here) 00:00 Krea 2 intro 00:30 Demos and aesthetics 02:04 How to use Krea 2 in ComfyUI 03:25 Downloading models 5:22 Krea 2 workflow 7:37 Luma Agents 9:41 How to use Krea 2 loras 12:30 How to make corn 14:54 How to train loras 15:22 Interesting details 17:52 License
19:09
aisearchio's profile picture

Krea 2 full review & installation tutorial (I'll be uploading more full videos on X. Please support me here) 00:00 Krea 2 intro 00:30 Demos and aesthetics 02:04 How to use Krea 2 in ComfyUI 03:25 Downloading models 5:22 Krea 2 workflow 7:37 Luma Agents 9:41 How to use Krea 2 loras 12:30 How to make corn 14:54 How to train loras 15:22 Interesting details 17:52 License

AI Search

30,091 views • 3 months ago