WaveSpeedAI Launches 'Desktop' for AI Power Users
WaveSpeedAI has launched WaveSpeedAI Desktop, a professional workspace designed for users who frequently run AI models. The platform aims to streamline AI generation tasks into a more collaborative and efficient workflow. The launch addresses the growing need for specialized tools to manage complex AI model generation and experimentation.
- WaveSpeedAI was founded in 2025 by CEO Cheng Zeyi and CTO David Li and is headquartered in Singapore with a base in Hangzhou, China. The company focuses on accelerating multimodal AI generation and secured multi-million dollar angel funding in April 2025 to optimize its inference engine. - The Desktop platform operates entirely locally on a user's machine, supporting Windows, macOS, and Linux, which ensures privacy by not requiring cloud uploads for processing. All generated assets are saved directly to the local drive, allowing for offline work and providing a version control history system for traceability. - Key features of the workspace include a multi-tab interface for comparing results from different models simultaneously and the ability to batch-process up to 16 variations at once for rapid iteration. - The platform allows users to save complex model and parameter configurations as "one-click templates" for instant reuse in future projects. It also offers full support for LoRA (Low-Rank Adaptation) for precise stylistic control over AI-generated content. - In addition to model workflows, the Desktop application includes several free, locally-run tools such as video upscaling, image resolution enhancement, background removal, and object erasing that function without subscriptions or internet access. - The company's core technology is a proprietary "in-house inference accelerator" designed to provide lossless speedup for AI-powered video and image generation. Its flagship image generation model, HiDream-I1, features 17 billion parameters.
