HackerTrans
トップ新着トレンドコメント過去質問紹介求人

jackson_mile

no profile record

投稿

Title: Show HN: Kling 2.6 Motion Control UI – Puppeteer static images with video

laike.ai
1 ポイント·投稿者 jackson_mile·7 か月前·1 コメント

Show HN: GPT Image 1.5 – An AI image editor with conversational editing

gptimage15.app
2 ポイント·投稿者 jackson_mile·7 か月前·0 コメント

Show HN: I collected and created some wallpapers for iOS 26

ios26wallpaper.org
1 ポイント·投稿者 jackson_mile·10 か月前·1 コメント

Nano Banana 3D Figurine

flash-image.art
2 ポイント·投稿者 jackson_mile·10 か月前·1 コメント

Show HN: I built a web UI for Gemini 2.5 Flash to solve AI image consistency

flash-image.art
2 ポイント·投稿者 jackson_mile·11 か月前·0 コメント

コメント

jackson_mile
·2 か月前·議論
To be honest, I think the performance of Gemini Omni Flash is still not as good as Seedance 2.0. You can try using both models on this platform. https://omnivideoai.co
jackson_mile
·3 か月前·議論
[dead]
jackson_mile
·7 か月前·議論
Hi HN,

I built a dedicated landing page and tool to utilize the new Kling 2.6 Motion Control API.

The main problem with existing AI video tools is the "stochastic" nature of movement—you prompt "running," and the AI guesses the physics. Kling's new update allows you to upload a reference video (e.g., a person doing a backflip) and apply that exact motion vector to a target image (e.g., a robot).

I wanted a cleaner interface to experiment with this specific "acting transfer" workflow without the clutter of the main generation platforms.

Key features enabled in this tool:

Support for full 30s continuous generation (no stitching).

Handling of complex limb occlusions (hands crossing bodies).

Simplified parameter tuning for "Creativity Strength".

Feel free to break it and let me know what you think.
jackson_mile
·10 か月前·議論
Hi HN,

I wanted to share a project I've been building to solve a problem I've personally faced.

As an indie dev and tabletop gamer, I've always found creating custom 3D character models to be a major hurdle. The learning curve for tools like Blender or ZBrush is steep, and while recent text-to-3D AI is promising, the results often aren't quite ready for a game engine or 3D printer without significant cleanup.

My goal with this project was to build a simple web tool that could take a detailed text prompt and generate a usable 3D asset. You can export models as OBJ/FBX for game development or as STL for printing miniatures.

The name is a bit of a nod to the core idea: "Nano" for the aim of achieving precise detail, and "Banana" for the goal of making the process as straightforward as possible.

It's still early, and the quality can definitely be hit-or-miss depending on the prompt. I'm launching it as a paid tool from the get-go, hoping to build it sustainably. I'm here to be transparent about that and would genuinely appreciate any feedback on the concept, the output quality, or anything else. I'll be in the comments to answer questions.