Hinge video prompt: 30 seconds, in-app recording, and Sume's role

No, going by Hinge's own announcement, you cannot upload a finished file as a video prompt. Hinge's newsroom post of September 20, 2022 says people can record 30-second videos right from within the app, and that they can preview before publishing. That describes in-app recording and a 30-second limit. Sume does not publish to Hinge, so the useful role for Sume is a rehearsal step before you record, and this post is honest about where that role ends.

The honest summary is short: the platform defines the format, and an outside tool cannot replace it. What follows is the small piece of work that is useful.

What Hinge says

The announcement is an older newsroom post and not a current help page, so confirm the numbers in the app before you plan around them. It states what is listed in the table. It does not say how many video prompts a profile can hold, so this post does not either.

Because the source is a newsroom post from 2022, features may have changed since. Hinge's current help center has articles on editing prompts, and the search result we saw described a 30-second limit as well, but we cite only the page we read in full.

Keep the scope small when you plan content for a dating profile. A clip you rehearse and trim for another channel is a different asset from a prompt answer recorded in the app, and mixing the two up wastes time on both.

Rehearse with a transcript

A prompt is a short answer to a short question. The fastest way to find out whether your answer fits in 30 seconds is to say it out loud and time it. If you record a rehearsal on your phone, you can have Sume probe and transcribe it. Video inspect with transcribe true runs Sume speech to text on the audio, at a public rate of $0.01 per audio minute, and returns a transcript with word-level data.

The sample below reads the word count of a transcript and the duration of the clip, then reports the pace. The 30-second figure comes from Hinge's page. Any pace thresholds are your own preferences, so the code reports the number and does not judge it.

A planning note about pace. Speaking rates differ a lot between people and languages, so there is no single correct word count for 30 seconds. Use your own rehearsals as the guide. Record three takes, compare the words per minute from the helper, and pick the take that sounds natural and fits.

def pace(transcript_text, duration_seconds, limit=30.0):
    words = len(transcript_text.split())
    wpm = words / duration_seconds * 60
    return {
        "words": words,
        "seconds": duration_seconds,
        "words_per_minute": round(wpm, 1),
        "within_limit": duration_seconds <= limit,
    }

sample = "I travel for the food, and I will always split the last dumpling."
print(pace(sample, 6.2))

Check for audio, and think about privacy

The inspect docs say a silent clip returns inspect_source_has_no_audio when you ask for a transcript, so probe has_audio first with frames set to false. Also, the rehearsal file is probably personal. It is hosted under your workspace on media.sume.com, so think about whether you want a copy of a private recording stored there.

The line this post draws

Do not use Sume to make a synthetic version of yourself for a dating profile video. The page you are following describes a feature built around recording yourself in the app, and a profile is a place where other people rely on what they see. This post does not describe any workflow to do that, and Sume documents none for this purpose.

Trim a rehearsal to see what fits

Trim can help in one more way. If your rehearsal runs 38 seconds, trim a 30-second range from a take and watch it back to see which sentences you lose. It costs $0.02 per job. The real recording still has to be made in the app.

After a trim, probe the result to see exactly how long it is. The trim result also reports actual_start_seconds, which matters only if you chose keyframe precision.