This server turns a topic or script into a finished, narrated, subtitled short video by writing/preparing voice-over, sourcing stock footage, and rendering an MP4 — all local, with free Edge TTS voices and no LLM API key.
Render a full short video (make_short_video): script → MP4 with neural voice-over, stock footage per search term (Pexels/Pixabay/Wikimedia), word-highlight or classic subtitles, optional background music, portrait/landscape/square aspect, custom voice, rate/pitch, clip durations, and your own media files or URLs.
Generate a voice-over (generate_voiceover): text → MP3 plus an SRT subtitle file timed to the words, in 300+ Edge neural voices across 70+ languages.
Browse voices (list_voices): filter the free Edge neural voices by language prefix (e.g. en-GB, ar) and gender.
Preview stock footage (search_footage): see clips for a search term (source, size, duration, author, licence) without downloading.
Check setup (doctor): verify ffmpeg, footage providers, and output/cache folders.
Pro extras (per README, not in this schema): license_status/activate_license and batch rendering via make_short_videos with branding, hook titles, caption presets, progress bars, thumbnails and extra aspect ratios.
Enables searching and downloading Pexels stock footage for video scenes, including clip previews and author/licence metadata for credits.
Enables searching and downloading Pixabay stock footage for video scenes, including clip previews and author/licence metadata for credits.
Enables keyless searching and downloading of Wikimedia Commons footage for video scenes, including author/licence metadata for credits.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@shorts-studio make a short about why honey never spoils"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Type a topic and get a finished vertical video: a script written by Claude, a natural neural voice-over, matching stock footage, word-by-word highlighted subtitles and a 1080×1920 MP4 ready for YouTube Shorts, TikTok and Instagram Reels.
/shorts-studio:make why honey never spoils
Real output: one command, one finished 1080×1920 video. More frames
No LLM API key. Claude is the scriptwriter, so there's no OpenAI or DeepSeek bill.
Free voice-over: Microsoft Edge neural voices (300+ voices, 70+ languages, including Arabic), with word-level timing.
Footage: Pexels or Pixabay (free keys), or Wikimedia Commons with no key at all. Falls back to Ken Burns zoom on photos. Letterbox bars are auto-cropped.
Viral-style subtitles: current-word highlight or classic lines; bottom, center or top; any colour, font or size; RTL-safe.
Your own media: your clips, images, voice recording and background music.
Credits file: every clip's author, licence and source link, ready to paste into the description.
Formats: portrait 9:16, landscape 16:9 or square 1:1.
Tested: 16 automated tests (unit, offline render, live TTS and live footage), Windows, macOS and Linux.
claude plugin marketplace add alycheaib/shorts-studio
claude plugin install shorts-studio@shorts-studioYou also need ffmpeg: winget install Gyan.FFmpeg, brew install ffmpeg or apt install ffmpeg. Or set FFMPEG_PATH.
Then, in Claude Code:
Command | What it does |
| One finished video, plus a title, description and hashtags |
| A series of videos and a posting plan |
Videos are saved to ~/Videos/shorts-studio/<date>-<title>/: video.mp4, voice.mp3, subtitles.srt, script.txt, credits.txt.
{ "mcpServers": { "shorts-studio": { "command": "npx", "args": ["-y", "github:alycheaib/shorts-studio"] } } }Related MCP server: Video MCP
Tool | |
| script and search terms → MP4 (voice, footage, subtitles, music) |
| text → MP3 and word-timed SRT |
| browse the Edge neural voices by language and gender |
| preview stock clips with author and licence |
| check ffmpeg, providers and folders |
| check or activate Pro |
| Pro: batch-render up to 20 videos with shared defaults |
Everything above is free forever, with no watermark. Pro is for creators and agencies who post every day under a brand:
Pro option | What you get |
| your logo on every frame (position, size, opacity), brand colour, custom font file, "Follow @handle" end card |
| a big title card for the first 3 seconds to stop the scroll |
|
|
| an animated watch-progress bar in your colour |
| one request → 9:16 and 1:1 / 16:9 from the same cut |
|
|
| a week of content in one command |
Ask Claude: "Make 5 shorts about saving money, with my logo C:\brand\logo.png, colour #00C853, handle @mybrand, hook titles, pop captions, progress bar, in portrait and square."
Plan | Price | |
Pro | $29 one-time | 1 person, all Pro options, free updates for v0.x |
Agency | $99 one-time | up to 5 seats, priority feature requests |
Get it: request a license and you'll get payment options (card, PayPal, USDT, Wise) and your key within 24 hours. Then tell Claude "activate my shorts-studio license SS1.…", or set SHORTS_STUDIO_LICENSE.
Keys are verified offline (Ed25519 signature): no account, no server and no tracking. Your videos never leave your machine.
Env var | |
| free at pexels.com/api, the best footage |
| free at pixabay.com/api/docs |
| a folder of royalty-free music; a random track is mixed in at 15% volume |
| default |
| a custom ffmpeg binary |
| extra fonts for subtitles |
| Pro license key (or use |
topic ──► Claude writes hook-first script + English search terms
──► Edge TTS (word boundaries) ──► captions (SRT + styled ASS)
──► Pexels / Pixabay / Wikimedia search ──► download + validate + cache
──► ffmpeg: cut 4 s segments, auto-crop, fill 9:16, concat
──► mix voice (+ music), burn subtitles ──► MP4 (H.264/AAC, faststart)npm install
npm test # 21 tests (the online ones skip without network)
npm run build # bundles plugin/server.mjsThe pipeline design is inspired by MoneyPrinterTurbo (MIT), rewritten from scratch in Node.js as an MCP server so it runs inside Claude. Respect each footage licence; credits.txt lists them for you.
Branded templates, auto-posting, a hosted API, or a white-label version for your agency: open an issue or contact Aly Cheaib.
MIT © Aly Cheaib
Check ffmpeg, footage providers, output and cache folders.
Inputs: no named parameters.
Free neural text-to-speech (Microsoft Edge voices, 300+ voices, 70+ languages) to MP3, plus an SRT subtitle file timed to the words.
Inputs: text (required), rate, pitch, voice, +2 more.
List the free Edge neural voices, optionally filtered by language prefix (e.g. "ar", "en-GB") and gender.
Inputs: gender, language.
Render a narrated, subtitled short video from a script you wrote: free neural voice-over (Microsoft Edge TTS), stock footage for each search term (Pexels/Pixabay with free keys, else Wikimedia Commons), word-highlight subtitles, optional background music, MP4 output. Takes 1-5 minutes.
Inputs: script (required), bgm, rate, order, +16 more.
Preview stock clips for a search term (no download): source, size, duration, author and licence.
Inputs: term (required), limit, aspect, provider.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 5 tools
Each tool has a largely distinct role: doctor is health-check, list_voices is discovery, search_footage is preview, generate_voiceover is an audio artifact, and make_short_video is the full pipeline. The only soft overlap is that make_short_video internally performs voiceover and footage steps, so an agent must infer that the single-purpose tools are building blocks or previews rather than alternatives — the descriptions clarify this, but only after reading.
Four of five names follow a clean verb_noun pattern (make_short_video, generate_voiceover, list_voices, search_footage). 'doctor' is a bare noun that breaks the convention, though it is a recognizable idiom for a diagnostics command.
Five tools is well-scoped for a video-generation server: one diagnostics, one orchestrator, and three focused helpers. Nothing feels padded or redundant.
The core lifecycle — check environment, pick voice, generate audio, scout footage, render video — is covered with no true dead ends. Minor gaps: search_footage only previews with no standalone download tool, and there is no way to list or supply background music, validate a script, or generate audio from an existing file.
Make, render and auto-post AI faceless videos to YouTube, TikTok and Instagram from any MCP client.
MCP server for Clipkit — gives AI agents a video toolbox via the Clipkit schema.
MCP server for Hailuo (MiniMax) AI video generation
MCP server for Google Veo AI video generation
