Agenvas is an open-source, self-hosted workspace for creating with AI. Its infinite canvas brings text, images, video, audio and AI agents together, allowing users to generate, edit, organize and reuse assets without moving between separate tools. Cards can be connected, arranged and linked to shared project context, while media proposals from agents remain subject to user approval.

The platform supports text generation through OpenAI-compatible endpoints and offers image, video and audio capabilities through administrator-published models, adapters and workflows. Integrations include ComfyUI, RunningHub and supported media services such as GPT Image, Google Nano Banana, Volcano Ark Seedance and Seed Audio. Image tools include annotation, cropping, rotation, mirroring, upscaling and depth extraction, where configured capabilities allow.

Agenvas also provides asset libraries, prompt templates, version history, task records and project export. Administrators can configure models, credentials and local or OSS, COS and S3 storage. Deployment uses Docker Compose and a single command, with no local build toolchain or GPU required.