nodetool-ai/nodetool

What it solves

NodeTool is an agent-first AI production studio designed for filmmakers and creators. It replaces the fragmented process of using multiple AI tools for scripting, storyboarding, and editing by providing a unified, local-first environment where an autonomous agent can handle the heavy lifting of production—from drafting scripts to cutting a multi-track timeline—while leaving the final creative control to the user.

How it works

The system is built around a node-based canvas where users can chain AI models and tools into workflows. An autonomous agent has full access to the studio's interfaces (node canvas, timeline, sketch pad, and script editor) as a toolbelt. This allows the agent to build workflows, generate footage, and assemble projects. Users can run inference locally using Ollama or MLX, or connect via API keys to cloud providers like OpenAI, Anthropic, and Gemini.

Who it’s for

It is primarily for filmmakers, video creators, and AI artists who want a professional-grade, editable production pipeline rather than a "black box" AI video generator.

Highlights

• Agent-Driven Production: An agent can pitch a concept, write a screenplay, storyboard scenes, and assemble a multi-track video project.
• Multi-Track Video Editor: Provides a real timeline for sequencing and compositing clips, allowing for in-place regeneration of AI footage without losing edits.
• Integrated Creative Suite: Includes a dedicated script editor with voice casting, a storyboard tool for low-cost pre-visualization, and a layered sketch editor for painting and generating.
• Local-First & Open: Projects and API keys stay local; the software is open-source (AGPL-3.0) and supports local inference.
• Extensible Architecture: Supports custom nodes in TypeScript or Python, JS scripts for data transformation, and an MCP server to connect with external agents like Claude Desktop.