AI-driven entertainment startup Dashverse secured $13 million in Series A funding. Peak XV Partners led the round. Existing investors Z47 and Stellaris Venture Partners also participated.
Powering Mobile-First AI Content Creation
Founded in 2023 by Sanidhya Narain, Lalith Gudipati, and Soumyadeep Mukherjee, Dashverse initially launched Dashtoon, an AI-powered digital comics platform. The Bengaluru- and San Francisco-based firm has since expanded its offerings. It now includes Frameo.AI, a generative video studio, and DashReels, a microdrama app.
DashReels garnered over 5 million downloads within one month of its launch.
The company's apps collectively boast a global user base exceeding 20 million.
Dashverse leverages generative AI to address high production costs for mobile-native episodic formats. This approach removes the need for complex tools and studio resources. Competitors like RunwayML and Midjourney also operate in the generative content space. The company aims to build infrastructure for the evolving creator economy.
Their technology, Hunyuan Keyframe LoRA, is an open-source system for AI video generation that lets creators define specific start and end frames—known as keyframes—and automatically generate smooth, coherent sequences between them. Built on the Hunyuan Video framework, it enhances traditional keyframe-based methods with architectural upgrades such as expanded input patch embeddings for better keyframe integration, LoRA-based fine-tuning to improve efficiency without retraining entire models, and conditioning mechanisms that ensure generated frames closely match the user’s visual anchors. Trained on a curated mix of high-quality public and proprietary video clips, the model uses motion-aware keyframe sampling to avoid artifacts and maintain temporal consistency. During use, creators can control keyframes, set variable video lengths, and optionally add descriptive text prompts for richer detail—making it a flexible, open alternative to proprietary tools like Runway’s Gen-3 or Kling, with full access to model weights and training code for community-driven innovation.
