Monet

Product Introduction

1. Definition: Monet is an agent-native, local-first desktop video editor and design canvas specifically architected for macOS. It functions as a specialized Integrated Development Environment (IDE) for video production, combining a multi-track timeline editor with a programmable design surface and an embedded pseudo-terminal (PTY) designed to host coding agents like Claude Code and OpenAI Codex.

2. Core Value Proposition: Monet exists to solve the "last mile" problem in AI-driven media creation by providing coding agents with direct, deterministic access to a video project’s state. Unlike traditional editors that rely on manual UI interaction or fragile plugin wrappers, Monet exposes its entire project model—including assets, sequences, and effects—as a structured graph. This allows AI agents to perform complex editing tasks, generate programmatic animations, and manage local media assets via a CLI-native workflow, effectively turning video editing into a software engineering task.

Main Features

1. Agent-Native Terminal and editorctl: Monet features a built-in terminal emulator powered by xterm.js and node-pty, allowing developers to run Claude Code, Codex, or custom scripts directly within the application context. The "editorctl" CLI acts as a local API bridge, providing a deterministic interface for agents to inspect the timeline, trim clips, add transitions, and manage the project graph without manual user intervention. This integration supports the Model Context Protocol (MCP), enabling seamless tool-calling for AI models to manipulate the video environment.

2. Monet Canvas (Design Surface): This is an artboard-based design environment integrated into the video editor. It supports three distinct modes: raw HTML/CSS/JS, Paper.js for vector graphics scripting, and Matter.js for 2D rigid-body physics simulations. Agents can programmatically generate visual assets—such as kinetic typography or physics-based animations—which are rendered as frames and automatically imported into the video editor’s media library for immediate use on the timeline.

3. Local-First AI Stack and Transcription: For maximum privacy and performance, Monet utilizes a local AI stack. It features an integrated faster-whisper runtime for on-device audio transcription and caption generation. While it utilizes OpenAI for embeddings and semantic search functionality, the project files (stored as .aiveproj.json) and media assets remain on the local file system. This architecture ensures that sensitive project data is not uploaded to third-party clouds for basic editing operations.

4. React-Based Video Composition with Remotion: The application leverages Remotion for its rendering engine and composition logic. This allows users and agents to create complex video components—such as lower thirds, audio visualizers, and title cards—using React code. These components are treated as standard timeline assets, enabling a hybrid workflow where programmatic elements and raw video footage coexist in a single, high-performance export pipeline (supporting 720p, 1080p, and 4K).

Problems Solved

1. Pain Point: The "Human-in-the-Loop" Bottleneck. Traditional professional video editors (e.g., Premiere Pro, DaVinci Resolve) are designed for manual mouse-and-keyboard input, making them inaccessible to autonomous AI agents. Monet eliminates this friction by providing a tool-surface that agents can "read" and "write" to directly.

2. Target Audience: Monet is designed for "Developer-Creators," AI engineers building automated content pipelines, technical marketers who need programmatic video generation, and power users of Claude Code or Codex who want to extend their coding capabilities into the multi-media domain.

3. Use Cases:

• Automated Video Summarization: Agents can scan local transcriptions, identify key segments via semantic search, and automatically assemble a rough cut.
• Programmatic Motion Graphics: Using the Monet Canvas to generate 100+ variations of a physics-based logo animation via a single Claude Code prompt.
• High-Efficiency Captioning: Generating and styling burnt-in captions locally using faster-whisper and Remotion-based text layers.

Unique Advantages

1. Differentiation: Most "AI video editors" are high-level SaaS wrappers around simple prompts. Monet is a professional-grade desktop application that treats video as code. It differentiates itself through its "Project Graph" model, where every clip and effect is a manipulable data point, rather than an opaque binary file.

2. Key Innovation: The integration of a PTY terminal and the editorctl bridge. By making the editor "agent-aware" at the kernel level, Monet allows AI to perform "surgical" edits (e.g., "split all clips where the word 'API' is mentioned and add a blue overlay") that would take a human editor hours to perform manually.

Frequently Asked Questions (FAQ)

1. Can I use Claude Code to edit videos in Monet? Yes. Monet includes an embedded terminal specifically designed to run Claude Code. By using the included editorctl CLI and MCP server, Claude Code can inspect your timeline, import media assets, perform cuts, and trigger exports directly from the command line interface within the app.

2. Does Monet require a cloud subscription for video rendering? No. Monet is a local-first application. Video rendering and composition are handled on your macOS device using Remotion and FFmpeg. While an OpenAI API key is recommended for semantic search and embeddings, the core editing, transcription (via faster-whisper), and export processes happen entirely on your local hardware.

3. What design tools are supported within the Monet Canvas? The Monet Canvas supports Paper.js for vector-based generative art, Matter.js for 2D physics engines, and standard Web technologies (HTML/CSS/JavaScript). This allows agents to write scripts that generate complex animations and graphics which are then automatically converted into video assets for your project.