Dreamspace

What is Dreamspace?

Dreamspace is an infinite canvas designed for exploring and interacting with the outputs of large language models (LLMs). It provides a unique visual environment where users can run prompts, compare outputs, and chain them together in a non-linear fashion. The core functionality revolves around a clickable canvas: clicking anywhere places a generation of the currently selected prompt, while clicking on other prompts changes the selection. Each text or image node can be clicked to reveal its details, including metadata and full content. Links between nodes visually represent relationships, such as chat message threads or prompt version histories. Users can continue from any message in a conversation, fork from any prompt to experiment with different text or models, and use previous outputs as context for new prompts. This makes Dreamspace ideal for researchers, developers, and creative professionals who need to iterate rapidly on AI-generated content, compare model behaviors, or build complex workflows. Key benefits include enhanced visualization of LLM outputs, seamless branching and versioning, and the ability to maintain context across multiple interactions. Use cases range from debugging model responses and exploring creative writing variations to designing conversational agents and analyzing prompt engineering strategies. Technically, Dreamspace supports multiple LLM integrations, real-time rendering of text and images, and persistent storage of canvas states. Its intuitive interface lowers the barrier for non-technical users while offering deep customization for advanced users. By transforming linear chat interactions into a spatial exploration, Dreamspace unlocks new possibilities for understanding and harnessing the power of large language models.

Key Features

infinite canvas prompt generation output comparison non linear chaining visual node linking conversation branching context reuse metadata display version history

Who is it for?

AI researchers, developers, creative professionals, prompt engineers, data scientists, product designers