Doubao Work Launches Canvas Feature and Upgrades Multimodal AI Models
ByteDance's Doubao Work has introduced an interactive infinite canvas feature designed for complex task management, allowing users to compare and edit text, layout, and visual elements on a single workspace. Additionally, the platform integrated the new Doubao 2.1 Lite model for high-frequency productivity tasks alongside the Seedream 5.0 Flash image generation model. This update aligns Doubao Work with mainstream global AI workspaces such as ChatGPT Canvas and Anthropic Artifacts by shifting from linear chat streams to flexible visual collaboration. The integration of lighter, faster models helps lower operational costs and token usage while improving efficiency for document writing, data processing, and design workflows. The canvas feature enables users to arrange raw materials, proposals, and generated outputs on one screen for granular, step-by-step adjustments to design, styling, and text. Doubao 2.1 Lite prioritizes faster delivery speed and lower quota usage for frequent use cases like documents, spreadsheets, and presentation slides.
## BACKGROUND
Canvas interfaces have emerged as a common pattern in AI products, enabling side-by-side editing of generated content instead of forcing users to work solely within a standard chat queue. ByteDance's Doubao suite relies on its proprietary large language models and the Seedream family of image generation models to power both consumer and enterprise AI applications.