shortshort

Purpose

shortshort is an AI clip maker for long videos that turns a long talk or podcast into 9:16 vertical short videos (shorts) through a web app. Its goal is to automate the process from topic selection, trimming, and framing to subtitles and publishing, and it is suited to working with long-form video content that already exists.

Main Capabilities

• Automatic clip finding: shortshort listens to the entire video and keeps complete, quotable passages. It selects content sentence by sentence and checks whether the end of each short video lands on a complete sentence, so the finished clip ends near the target duration instead of cutting a sentence in half.
• Face tracking and reframing: each clip is cropped to 9:16, and the frame follows the most prominent face on screen and moves with it, with automatic zoom (about x1.0 to x1.6). Slides, screens, and wide shots remain full-frame. The official explanation says it does not determine who is speaking, so in two-person shots the crop may remain on the listener, and there is no manual framing control.
• Word-by-word subtitles: every word is aligned to the audio timing. Four subtitle styles are offered: Spotlight (default, uppercase and highlights the word currently being spoken), Impact (highlights one bold word per line), Karaoke (keeps the whole line and lights up word by word), and Minimal (clean, readable text).
• Scheduled publishing: you can connect a YouTube channel, choose a date and time to publish each short video, and titles and descriptions are generated in the selected language, or shortshort can automatically fill the schedule. TikTok publishing is in early access. Instagram is not integrated, so you need to download the MP4 and publish it yourself.
• Multilingual (follows the video language): transcription, clip selection, and subtitles follow the language spoken in the video, and subtitles are not translated. Transcription uses OpenAI's whisper-1, which the official description says supports 98 languages. The interface and AI-suggested titles are in English, and you can rename them yourself.

How to Use

The core workflow described officially is:

1. Upload: a single file up to 2GB, supporting MP4, MOV, and WebM.
2. Find clips: listen to every sentence.
3. Trim: keep complete passages.
4. Reframe: follow the speaker in 9:16.
5. Add subtitles: word-by-word timing.
6. Schedule: choose a date.
7. Publish: automatically generate a title and publish.

Users need to choose the desired number of short videos and the duration of each short video. Official example: a 2:04:55, a 1920×1080 MP4 video, generating 3 30-second short videos (about 1:30 target duration), consumes about 125 credits. Official documentation states 1 minute of video consumes 1 credit. When using an account, analysis runs on its servers, and the tab can be closed after upload completes.

A free account is required, initially providing 30 credits, no credit card needed. It handles running AI analysis, opening the editor, and exporting MP4. Each short video is an AI-generated suggestion, and clips, titles, descriptions, and subtitles can be edited before export.

Use cases

Official documentation lists it as suitable for existing videos such as podcasts, webinars, courses, talks, and live streams.

Limitations and notes

• A single uploaded file is at most 2GB; how many hours it can hold depends on the bitrate. Recordings over 2GB can be re-exported after lowering the bitrate or resolution.
• The free plan allows at most 3 short videos per video, and exports carry a small watermark; paid plans remove the watermark and allow 5, 10, or 20 short videos per video.
• The target duration of each short video can be selected between 15 and 90 seconds, and the actual duration may be up to 25% shorter or longer than the target.
• Export formats: 9:16 1080×1920 MP4, or 4:5 1080×1350 reframed for the Instagram feed.
• Publishing: only YouTube supports automatic publishing; TikTok is in early access; Instagram is not connected.
• Data storage: source videos, short videos, and account data are stored by Supabase in the EU (Frankfurt), and shortshort operates in France. AI analysis is the exception, where videos or transcripts are sent to Google Gemini or OpenAI (outside the EU), and Google files are deleted when the task ends. Scheduled short videos are uploaded to YouTube.
• Visibility: uploads and drafts remain private before export. Exported short videos are published by default to the community gallery, including the full source video and the transcript of that short video. Any account can withdraw a published short video at any time, and paid plans have a setting to keep all short videos out of the gallery.

Not yet provided

Official materials do not provide specific pricing, paid plan names and amounts, the number of credits corresponding to each plan, subscription periods, or other billing details; the launch time of the TikTok publishing feature is also not stated.