Video editing has long been one of the most manual and tedious processes in creative production. While generative AI has already revolutionized the creation of individual clips, the transition from 'generation' to 'structured editing' remained a bottleneck, requiring users to act as intermediaries between the prompt and the timeline. Palmier Pro fundamentally changes this paradigm, positioning itself not just as an editing tool, but as an operating environment where users and AI agents collaborate in real-time on the same project.

A Native NLE Built for the Agent Era

Developed entirely in Swift, Palmier Pro is not a simple wrapper of existing tools, but a macOS-native video editor (specifically optimized for Apple Silicon and requiring macOS 26 Tahoe) that takes inspiration from the Premiere Pro workflow. However, its true innovation lies in the deep integration with AI agents.

According to the official GitHub repository, the software allows models like Claude, Codex, or development tools such as Cursor to interact directly with the timeline. This is not about simple text suggestions, but concrete operationality: agents can perform trimming, reorder sequences, and generate new footage without the user having to manually move every single element.

The Technical Core: MCP and LLM Integration

The pillar enabling this interaction is the Model Context Protocol (MCP). Palmier Pro exposes a local MCP server (at http://127.0.0.1:19789/mcp) that acts as a communication interface between the editor and external LLMs.

This architecture allows tools like Claude Code or Cursor to "see" the timeline state and send precise commands to manipulate the project. Essentially, the editor becomes an extension of the AI agent: a professional can ask Claude to "shorten all clips in the second sequence to make them more dynamic," and the AI will physically execute that operation within the software. Technically, the system relies on AVFoundation to ensure high performance during composition and playback, utilizing a proprietary file-package format (.palmier).

Integrated Generative AI and Business Model

Beyond editing automation, Palmier Pro integrates asynchronous pipelines for asset generation. Within the timeline, users can employ SOTA models such as Seedance, Kling, and Nano Banana Pro to create video and image assets on the fly.

The project adopts a hybrid distribution model: the video editor, the MCP server, and the agent chat are fully open source under the GPLv3 license. This means anyone can download and use the software as a traditional NLE (similar to CapCut or Premiere) for free and without requiring an account. The only closed-source component is the Generative AI processing, which requires a credit-based subscription. According to details from eesel AI, plans start at $29 per month (5,000 credits) up to $69 (12,000 credits), while editing and exporting remain free.

Toward a New Way of Content Production

The approach taken by Palmier Pro fits into a broader trend of "LLM-programmable interfaces." Just as we recently saw Unreal Engine open its doors to LLMs to revolutionize game design via prompts, Palmier Pro applies the same logic to video editing.

We are witnessing a shift from "software as a tool" to "software as a collaborator." The ability to delegate repetitive and technical tasks (such as clip reordering or basic synchronization) to an AI agent allows the creator to focus exclusively on creative and narrative direction. Although current limitations to macOS 26 Tahoe may hinder immediate mass adoption, the open-source nature of the editor's core could accelerate the development of plugins and integrations that make AI-assisted editing an industry standard in the coming years.