Levea

Agentic Video Editor

MCP server17 viewslaunched Oct 2, 2026

Upvoted by 1 builder

Comments

About Levea

Levea is an agentic AI video editor that turns natural language direction into finished, professional vertical and horizontal videos—without spending hours manually trimming clips and syncing subtitles.

Key features: • Prompt-First Editing: Describe what you want in plain words. Levea plans, cuts, and sequences your video while leaving you in full control of an editable timeline. • Kinetic Typography: 40+ high-end caption styles with word-by-word active animation and crisp SDF text rendering. • Smart Perception: Active speaker face tracking, automatic vertical 9:16 reframing, silence/filler-word removal, and split-screen layouts. • Verified Motion Graphics: HyperFrames kinetic overlays, lower-thirds, tweet/reddit callouts, and animated data cards. • Multi-Surface Integration: Edit directly in the browser studio, on Telegram, or inside Claude Code and Cursor using our open-source MCP server (`npx -y levea-mcp-server`).

Overview

Who is it for?

Content creators, video editors, marketing agencies, and developers who produce short-form videos (TikTok, Instagram Reels, YouTube Shorts) and want to automate trimming, vertical reframing, and subtitling. Also built for AI agent builders using Claude Code, Cursor, or OpenClaw who want to edit videos directly from their developer environment.

Problem

Manual video editing is tedious and slow. Trimming footage, finding viral hooks, aligning subtitles word-by-word, reframing 16:9 video to 9:16 vertical, and keyframing motion graphics takes hours per video. Traditional one-shot generative AI models output locked, un-editable pixels that cannot be tweaked or fixed.

Solution

Levea operates on the Prompt-to-Project paradigm: you provide a creative prompt or brief, and Levea's deterministic Workflow DAG engine automatically schedules trimming, Whisper speech transcription, active speaker face tracking, 40+ kinetic caption styles, and HyperFrames motion graphics—while maintaining an editable timeline you stay in full control of.

What makes it unique

Unlike standard AI tools that hallucinate edits in loose LLM loops or output un-editable video files, Levea uses Frontier LLMs strictly for semantic intent and compiles that intent into a deterministic Workflow DAG executed by native Vulkan and WebGPU pipelines. Real-time 60 FPS WebGPU preview in the browser matches native 4K cloud rendering frame-for-frame, and our open-source MCP client connects directly into Claude Code and Cursor.

The maker

brajendra kumar

Maker · @brajendrak00068

Details

Platforms

WebWindowsmacOSLinuxiOSCLIAndroidChrome extensionAPI

Share this launch