A D-ID Alternative When You Only Need One Talking Photo

D-ID does the same core trick this site does — a still photograph, a generated voice, a mouth that moves with the words — and it has been doing it longer, at a bigger scale, with an API behind it. The difference worth understanding is how each one charges.

D-ID's plans, as published

Prices checked on d-id.com on 17 September 2026:

• Trial — $0 for 14 days, 3 minutes of video in total, full-screen watermark, 1 personal avatar
• Lite — $4.70/month, billed $56 annually: 10 minutes a month, D-ID watermark stays
• Pro — $16/month, billed $191 annually: 15 minutes a month, premium voices, 1 voice clone, commercial licence, AI watermark
• Advanced — $108/month, billed $1,293 annually: 100 minutes a month
• Enterprise — quoted

Two things to notice. The entry paid plan still watermarks your video, so the first tier that gives you a clean file is $16 a month billed as $191 up front. And the allowance is minutes per month — a budget for continuous production, which is exactly right if you are making agent videos or translating a catalogue, and odd if you want one twelve-second clip.

What one clip costs here

$4.99, once. That is 50 credits: 2 talking photos at 20 credits each, or 5 videos on the other tools at 10 each. No subscription, credits good for a year, failed runs refunded to your balance automatically.

There is no free talking photo — every run pays for a generated voice and a generated video, so giving them away is not something a small site can do. The other tools (animate a photo, old family photos, make it dance) do include one free video per account, no card.

Where D-ID is clearly ahead

• An API, with real documentation and support behind it
• Longer videos — minutes, not seconds
• Voice cloning on Pro and above
• Agents — interactive avatars that hold a conversation
• Commercial licence spelled out per tier

If your use case is a product, a client deliverable, or anything that needs a licence attached, buy from the company that publishes one.

Where the small tool fits

A single photograph, a single sentence, once or twice a year. Birthday messages. A grandparent's portrait saying the thing they always said. A pet that finally speaks. A character for a classroom.

For those, the specifics: front-facing portrait, one person, mouth visible; up to 180 characters of script, which is roughly twelve seconds; four voices; 480p output; two to five minutes to generate.

The test

Add up what you would make in a year. Under about ten clips, the pay-once tool is cheaper and there is nothing to cancel. Above that, or if you need an API, minutes-per-month pricing is the right shape and D-ID is a good version of it.