Edit Pro for YouTube: AI video editor for Spanish talking videos

Description

Edit Pro for YouTube is an automatic editor for talking videos that runs on your computer. You give it a talking video in Spanish and it gives you back the clean video, with captions, animated graphics and leveled sound, in landscape, vertical or square.

Language: it only works with videos spoken in Spanish. Transcription is fixed to Spanish and filler-word detection is designed for Spanish. The documentation and the Chrome panel are in Spanish too.

It works for YouTube (16:9, with chapters) and, with the same video in 9:16, for Shorts, Reels and TikTok. Inside it is called CutPilot: that is the name you will see in the ZIP, the README and the install guide.

The problem it solves

Editing a to-camera video eats up time on repetitive tasks: finding the "um"s and the silences, cutting the sentences you repeated, adding captions, leveling the audio and making a vertical version. This studio does that work from the word-by-word transcript, so you only review.

How it works, in 3 steps

• Copy your recording into the entrada/ folder or drag it onto the Chrome panel.
• The studio cleans the audio, transcribes, removes filler words, silences and retakes, and Claude proposes the graphics, saying which word each one goes on.
• Review the contact sheet if you want to approve before rendering, and get the final video with its quality report.

What you get

• The editar.py orchestrator and the pipeline modules: noise reduction, transcription, cleanup, visual plan, face detection, rendering, mixing and QA.
• The Remotion project with word-by-word captions, data cards, lists, chapter titles, zooms, intro and outro.
• A skill and a reviewer subagent for Claude Code, plus a hook that protects your original videos.
• The Chrome extension with its local server: job queue, brand kit, settings and diagnostics.
• 55 CC0 sound effects (Kenney), 28 OFL-licensed fonts and tools to render offline and to import music you already downloaded (for example, from the YouTube Audio Library).
• An installer that downloads the models and DeepFilterNet from their official sources, with SHA-256 verification.
• Automated tests and a config.yaml file with every parameter commented.

Who it is for

Creators, teachers, coaches and marketing teams who record talking videos in Spanish and use Claude Code, and who prefer processing on their own computer instead of paying per-minute credits.

Requirements

• uv (it installs Python 3.12), Node.js LTS and ffmpeg 7 or newer.
• DeepFilterNet for noise reduction (optional: the installer downloads it; without it, ffmpeg's filter is used, with lower quality).
• Claude Code for the AI decisions (optional).
• An NVIDIA GPU is recommended; without one, Whisper runs on the CPU.
• About 20 GB free: the Whisper model downloads the first time (about 1.6 GB).

Honest limitations

• Tested only on Windows 10/11. macOS and Linux are not verified, and the Chrome extension registers itself automatically only on Windows.
• It edits the video you give it and delivers it in three formats: it does not cut several reels from one long recording, and it does not write titles or descriptions.
• Videos in Spanish only.
• Installation uses the terminal, and the extension loads in Chrome's developer mode.
• Rendering takes time: the measured estimate is 35 to 45 minutes per format for 10 minutes of raw footage.
• Whisper can mishear words in very short clips and an intentional repetition may get cut: check the .srt and the cut list in the report.
• Remotion has its own license; companies with more than 3 employees need the company license. Other dependencies installed separately have their own licenses (details in THIRD_PARTY_NOTICES.md).
• For music and brand logos, use only licensed material: the studio records the source and license of each asset, but you are responsible for how you use them.