Pet shelter monthly posts: 8 photo clips with captions, $5.72

A small shelter that posts two adoptable-pet clips a week can do it for $5.72 a month on Sume: eight MiniMax H3 Max clips at $0.50, eight caption jobs at $0.20 and one $0.125 music bed. The captions are the part to get right, because a clip with no speech cannot use the speech-to-text path.

Use the real photo of the real animal as the start image. The goal is movement around a true picture, not a generated pet that does not exist.

Why captions need cues here

Video captions normally transcribe speech. The docs state that a silent clip fails with caption_no_speech and a next_action of use_overlay_captions. The fix is to pass cues (or segments) with the text and timing you want, which skips speech-to-text. For a pet post that means a name, an age and one line, such as "Miso, 2 years, ask about her." The caption price is the same flat $0.20 per job for videos up to 60 seconds.

Choosing the per-clip model

H3 Max at $0.50 per clip is the plan's main line. If the images are strong, a cheaper model could carry some posts; the price ranking shows the options. Check the animal still looks like itself after motion: adopters will compare the video to the real animal.

Do not add a talking voice. Without speech you avoid the speech-to-text step and its failure mode, and the music bed stays under the captions.

Budgeting the wobble

Plan for about one retry in four on motion that distorts a face. Two extra clips add $1.00, so a monthly top-up of $8 covers the plan plus retries with room left.

Reserve at submit, capture on success, refund on failure: a failed clip does not consume the plan, but the hold must fit in the balance when you submit.

What the shelter should write in the captions

Captions earn their $0.20 only if they carry the information an adopter needs. Keep each cue to a few words: the name, the age, one trait, and where to ask. Authored cues let you control line breaks and timing, which speech captions cannot do for a silent clip.

The docs also note that estimated duration above 60 seconds is rejected for inline captions, and the pet clips here are far shorter, so the flat caption price applies.

• Cue 1: the animal's name and age.
• Cue 2: one specific trait that is true.
• Cue 3: how to ask, such as a phone number or the shelter's page.
• Check the spelling of every name before submitting; a recaption is another $0.20.

Scaling the plan

Doubling to four clips a week gives 16 clips, 16 caption jobs and one music bed: $11.32 a month. The music bed does not need to change monthly; one track can run under every clip. Photos are the real constraint. A clip built from a blurry photo will look blurry in motion, so spend an hour on pictures before you spend a dollar on clips.