Higgsfield AI launches Canvas, a node-based workspace for AI production

Higgsfield AI launches Canvas, a collaborative node-based workspace to build image and video pipelines with models like Kling 3.0, Wan 2.7, and Soul 2.0.

Higgsfield Canvas, a node-based workspace designed for team brainstorming, pre-production, and the creation of reproducible image and video generation pipelines. Each node represents a distinct element (prompt, image, visual reference, or AI model) that the user connects to orchestrate a complete workflow: generate a character, animate it, associate a product with it, then export a series of variants, without leaving the interface or writing code. The environment is collaborative in real-time, following the Figma model. A shared link allows multiple team members to edit the same canvas, add nodes, generate content, and communicate via an integrated chat. Versions are saved automatically, and each node can receive individual comments. Once built, a pipeline can be reused by modifying only the inputs, which is particularly aimed at marketing campaigns, storyboards, or VFX uses where the same logic is replayed on a large scale. The canvas centralizes proprietary and third-party models already accessible on the platform, including Soul 2.0, Seedance 2.0, Kling 3.0, Wan 2.7, Veo 3.1, GPT Image 2.0, and Nano Banana Pro, which can be combined within the same pipeline. An integrated agent, accessible in natural language, can build a complete workflow from a simple instruction, such as a request for a scalable product pipeline. The tool is available at higgsfield.ai/canvas, with free access limited by a credit system consumed upon generation, and paid plans starting from approximately 5 dollars per month annually for the Basic plan, up to a Business offer designed for teams.