TypeDance is an AI-assisted tool designed to create semantic typographic logos, which blend typeface and imagery to represent semantic concepts while maintaining legibility. The tool addresses the challenges of blending geometrically dissimilar typefaces with semantics and the lack of personalized design in end-to-end generative approaches. TypeDance incorporates design principles, supports flexible typeface selection at different granularities, and enables imagery mapping at various structural levels. It includes a comprehensive design workflow, including ideation, selection, generation, evaluation, and iteration. The tool leverages combinable design priors extracted from uploaded images and uses a diffusion model for blending. A user study confirmed the usability of TypeDance in generating diverse and expressive semantic typographic logos across different scenarios. Key contributions include a formative study identifying generalizable design patterns, an intent-aware input based on user-personalized images, a blending technique supporting flexible blending at different granularities, and an integrated authoring tool facilitating personalized design.TypeDance is an AI-assisted tool designed to create semantic typographic logos, which blend typeface and imagery to represent semantic concepts while maintaining legibility. The tool addresses the challenges of blending geometrically dissimilar typefaces with semantics and the lack of personalized design in end-to-end generative approaches. TypeDance incorporates design principles, supports flexible typeface selection at different granularities, and enables imagery mapping at various structural levels. It includes a comprehensive design workflow, including ideation, selection, generation, evaluation, and iteration. The tool leverages combinable design priors extracted from uploaded images and uses a diffusion model for blending. A user study confirmed the usability of TypeDance in generating diverse and expressive semantic typographic logos across different scenarios. Key contributions include a formative study identifying generalizable design patterns, an intent-aware input based on user-personalized images, a blending technique supporting flexible blending at different granularities, and an integrated authoring tool facilitating personalized design.