Multicam skill
A skill that turns one talking-head clip into a virtual multi-camera cut prompt for Google Omni.
- 64 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill reads a local video and, with your permission, installs dependencies on the machine.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/multicamDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill reads a local video and, with your permission, installs dependencies on the machine.
This is third-party code. Review the repository files before installing.
What it does
Multicam takes video shot on one camera and prepares a prompt that makes Google Omni re-frame the footage from new angles with hard cuts. The skill transcribes speech locally with Whisper at word level, finds pauses and phrase starts and proposes four cut points that avoid mid-word positions. The user confirms timestamps and angles: left profile, high angle, close-up and a return to the original frame. The skill then fills the frozen template, verifies it with a script and saves a paste-ready English prompt.
Who it is for. Short-video creators and social media specialists who shoot one take and want a multi-camera edit effect.
Good fit when
- A clip was shot on one camera but needs a punchy cut by phrase.
- Cut timestamps must follow speech pauses rather than be placed by eye.
- A finished prompt must be pasted into Google Omni together with the source clip.
- You want to confirm the prompt template stayed unchanged outside the four timestamps.
Not a fit when
- Classic timeline editing with real takes is needed.
- The video has no speech to place cuts against.
- Guaranteed audio and lip-sync preservation is required.
Example request
Make a multicam prompt for this video: find four cut points on speech pauses and show why they were chosen.Limitations
Google Omni does not always obey the prompt: the author measured that it can rewrite audio and drop words, shift cuts by up to 0.2 seconds and output 720p. After generation, re-transcribe the result, measure the cuts with ffmpeg and upscale; the recipe is in `SKILL.md`. The first transcription downloads a Whisper model of about 460 MB. It needs Python 3.8 or newer, ffmpeg and the `openai-whisper` package. Windows and Linux have their own ffmpeg install commands in the README.
How to disable. Delete the `~/.claude/skills/multicam` folder. The Whisper model in the user cache and finished prompts are removed separately.
FAQ
Are keys or paid services required?
No. Transcription runs locally through Whisper, with no keys or accounts. Only Google Omni access itself may cost money.
What does the preflight script check?
`scripts/check_env.py` reports which programs are missing and what they are for, and asks permission before installing.
How are cut points chosen?
`scripts/beats.py` finds phrase starts and breath pauses. Cuts land on a phrase start or inside a pause, never mid-word.
Can I change the angles?
Yes, there are presets: left profile, high angle, close-up and a return to the original frame. Any can be replaced with a description.
What language is the prompt in?
English. Whisper handles many languages, so the speech can be in any language.
Discussion
No questions or reviews yet.
Related
SkillsA skill that turns a topic into a narrated 4K video: research, script, local TTS, Remotion composition and an FFmpeg mix
A skill that turns a coding agent into a studio for voiceover explainer videos: word-level sync, a motion library and Remotion assembly
A skill that turns a messy phone recording into a vertical reel: best takes, snap zooms on cuts and word-timed captions.
A skill that pulls highlights from a long video: download, transcript, segmentation, analysis and clips with titles
The /watch skill turns a link or file into frames and captions, or with a Gemini key hands the whole video to Google for the agent to answer from
Skill for Codex and Gemini: turns a phrase, text or SRT into a Vox-style storyboard with prompts and JSON for Google Flow
README in short
The skill prepares a prompt with four cut timestamps based on real speech. The template is frozen and checked by a script.