#AI

Experimenting with Hyper3D Rodin's multi-image input to improve 3D generation quality — and how T-pose vs. standing pose changed the results.

Using Hyper3D Rodin to generate 3D models from text prompts and images — plus tracking down and fixing a bug encountered along the way.

How to have Claude AI operate Blender with MCP. Setup and basic usage.

Planning a Japanese voice chat system built on the voice-chat project, using Gemini 2.0 Flash for LLM and VOICEVOX for TTS.

A workaround for the bug where WebFetch(domain:*) wildcard doesn't work — achieving equivalent functionality with curl and a custom skill.

Thoughts on design tools, vibe coding, and team division after reading a survey report.

To stabilize body shape in Gemini Gem and Flow, I built a body reference image that includes figure information.

How to combine UX psychology, frontend-design SKILL.md, and the Atlassian Design System to break out of the generic AI-looking design trap.

Testing the image generation features of Flow, now available in Google AI Pro. Findings on the He/She pronoun issue, the effectiveness of natural English prompts, and how to use Flow vs Gem.

After reading shuntaka’s article, I reviewed my own Claude Code workflow: keep sessions vs git-push, document structure, and skill operations.

After reporting that Nano Banana Pro wasn't available even with a Gemini Pro subscription, it finally rolled out — here's what character-consistent image generation actually looks like in practice.

A record of designing prompts to make Gemini's image generation correctly draw a side ponytail, and building a Gem using a full 360° set of reference images.