MMCPMarkets
HP

hirogoing/polystudio

Developer tools

This page was translated automatically. Technical details may differ from the original. View the original

74 stars Quality 46 Trend 14
Repository Install

A "Conversational Image/Edit/Edit + Visual Artboard" project: FastAPI backend calls via LangGraph orchestration tool, React frontend provides artboard hosting and project management, supports SSE streaming output, automatic image dropping to artboard, project links, and historical records.

Overview

PolyStudio is a conversational multimodal content generation platform that combines language models with professional tools (such as image generation, video generation, 3D model generation, and multimodal content understanding), enabling the generation, analysis, and management of multimedia content through natural language dialogue. The project uses FastAPI + LangGraph to build an intelligent agent orchestration system, providing unlimited artboard hosting and project management, supporting SSE streaming output, automatic content insertion, project link sharing, and other functions. PolyStudio supports flexible API integration, allowing you to easily replace it with any image, video, or 3D generation service to create your own multimodal creation tool. It also supports writing custom skills through SKILL.md, expanding the agent's specialized domain capabilities. Supports integration with OpenClaw — after configuring polystudio-client/SKILL.md as a skill, external agents can directly create via natural language driving PolyStudio, with generated content automatically synced to the canvas. - : Drives PolyStudio via curl / HTTP, supports SSE streaming consumption, media URL extraction, multi-turn dialogue, allowing users to sync generated content to the canvas 🆕 even when not on the page - : Images/videos/3D models generated by external agents when users leave the canvas page are automatically inserted upon returning to the page without losing 🆕 - : Added paper-writing, Four out-of-the-box custom skills 🆕 — video-creator (long video generation), virtual-anchor (virtual human synthesis), and podcast-creator — inject cross-dialogue long-term memory into the Agent through AGENTS.md / TOOLS.md / IDENTITY.md / USER.md / SOUL.md / MEMORY.md, setting page visual editing - : a brand-new visual configuration center supporting visual configuration of tools, skills, MCP, environment variables, and workspaces - : by writing SKILL.

Overview (machine translated)

Python 3.9+ | Node.js 18+ | License: MIT

PolyStudio is a conversational multimodal content generation platform that combines language models with professional tools (such as image generation, video generation, 3D model generation, and multimodal content understanding), enabling the generation, analysis, and management of multimedia content through natural language dialogue. The project uses FastAPI + LangGraph to build an intelligent agent orchestration system, providing unlimited artboard support and project management, supporting SSE streaming output, automatic content insertion, project link sharing, and other functions. PolyStudio supports flexible API integration, allowing you to easily replace it with any image, video, or 3D generation service to create your own multimodal creation tool. It also supports writing custom skills through SKILL.md, expanding the Agent’s specialized domain capabilities. Supports integration with OpenClaw — after configuring polystudio-client/SKILL.md as a Skill, external Agents can directly create via natural language-driven PolyStudio, with generated content automatically synced to the canvas.

Feature Overview

  • External Agent Access (polystudio-client): Drives PolyStudio via curl / HTTP, supporting SSE streaming consumption, media URL extraction, and multi-turn conversations, allowing users to sync generated content to the canvas 🆕 even when not on the page
  • Screen Backend Media Synchronization: Images, videos, or 3D models generated by external agents when the user leaves the canvas page are automatically added and inserted upon returning, no longer lost 🆕
  • Built-in Skills Expansion: Four new out-of-the-box custom skills 🆕 have been added: paper-writing (paper writing), video-creator (long video generation), virtual-anchor (virtual human synthesis), and podcast-creator (podcast production).
  • Workspace Identity Memory: Injects cross-dialogue long-term memory into Agents via AGENTS.md / TOOLS.md / IDENTITY.md / USER.md / SOUL.md / MEMORY.md, enabling visual editing of settings pages
  • Settings Page: Brand new visual configuration center, supporting visual configuration of tools, skills, MCP, environment variables, and workspaces
  • Skill Extension System: Extends the agent’s specialized domain capabilities by writing SKILL.md files, supporting progressive loading on demand
  • Multimodal Content Understanding: Based on Qwen3-Omni-Flash, supports intelligent analysis of images, audio, and video, with dual-modal output of text + voice, and supports the integration of ‘understanding → creation.’
  • Video Upload and Management: The frontend supports uploading video files (MP4/MOV/AVI/MKV/WEBM), which are visualized in conversations and saved locally
  • Intelligent Podcast Generation: A complete audio content generation workflow from scripting to mix output

This page was translated automatically. Technical details may differ from the original.

Recommended Tools

Try a different keyword or remove a filter.

Install

npx skillfish add hirogoing/polystudio
New to MCP? Start with the mcpskill.org guide →