AvatarMotion with Caption was tested with one simple question: how much can a themed portrait-to-video preset do when the input stays simple and the caption stays short? The four clips below use the same basic setup – one clean, front-facing portrait, a 1-2 word caption, Standard video mode, and a fixed seed of 42. They show the birthday and zodiac options for both woman and man presets rather than a broad claim about every kind of avatar video.
Model and scope
AvatarMotion with Caption on Wiro turns one uploaded portrait into a themed short video and can return a video, an image, or both. The available effect choices are Woman Birthday Mode, Man Birthday Mode, Woman Zodiac Season, and Man Zodiac Season. Its model page identifies the workflow as image-to-video and lists Standard and Pro as the two video-mode choices.
This is a preset tool, not a freeform talking-head generator. The useful test is whether the selected template fits the person in the input and whether the caption remains readable while the effect plays. The underlying model listing carries a Wan2.2 tag. For technical background on that family, see the Wan2.2 image-to-video model card and the Wan2.2 GitHub repository.
Test setup
The inputs below are deliberately ordinary portrait photos: one woman and one man, each with the face clearly visible and the subject separated from a busy background. Each run used videoMode: standard, outputType: both, and seed: 42. Holding the seed and mode constant makes the preset choice and caption the variables that matter here.
That matching matters. The preset names explicitly separate woman and man variants, so the test did not mix them. A close crop, sunglasses, hair across the face, or strong side lighting would be a different test. The documentation recommends short promotional text and gives one or two words as the target. That is why the birthday caption is a number and the zodiac captions are single signs.
What the four outputs show
1. Woman Birthday Mode with caption 30
Parameters: effectType: birthdaymode-woman-birthday-mode; caption: 30; Standard mode; Video and Image output; seed 42.
This clip is the clearest fit for a birthday greeting because the animation, decorative treatment, and numeric caption point to the same occasion. The number is short enough that it reads as a label rather than a sentence. It also shows the practical boundary of the tool: it supplies a themed moment around the portrait, not a custom narrative or an editable scene. Pick this preset for a quick birthday post where the portrait is the main asset.
2. Man Birthday Mode with caption 30
Parameters: effectType: birthdaymode-man-birthday-mode; caption: 30; Standard mode; Video and Image output; seed 42.
The second birthday result checks whether the birthday idea still reads when the matching portrait and man-specific preset change. It does: the subject remains the focal point and the caption stays compact. The useful lesson is not that one preset is inherently better than the other. It is that a matching preset reduces avoidable mismatch. Use this option for the same short-form birthday use case with a male-presenting portrait, and keep the text to a number, name, or brief greeting.
3. Woman Zodiac Season with caption ARIES
Parameters: effectType: zodiacmode-woman-zodiac-season; caption: ARIES; Standard mode; Video and Image output; seed 42.
The woman zodiac output makes a different kind of promise: it is less about an event and more about an identity label. ARIES works because it is one word and visually distinct. The output demonstrates why the caption field should not carry a full marketing line. The preset already provides the visual context; adding a sentence would compete with it. Pick this version for horoscope content, a birthday-month series, or a lightweight profile post.
4. Man Zodiac Season with caption LEO
Parameters: effectType: zodiacmode-man-zodiac-season; caption: LEO; Standard mode; Video and Image output; seed 42.
The final clip repeats the zodiac test with the matching man preset and the label LEO. Like the woman zodiac result, it works best as a self-contained social asset. It does not prove that the system will preserve a face under every pose or lighting condition; these are clean portrait inputs. It does show that the preset can package a single portrait and a very short label into a ready-to-share themed video.
Runtime and cost per output
The saved four runs do not include task-detail records, so this test cannot honestly assign a measured duration or charge to each individual video. The model documentation does provide a reference task record: it reports elapsedseconds: 6.0000 and totalcost: 0.003510000000 for a documented example task. Treat that as a documentation example, not a promise for these four outputs. Queue time, selected mode, file handling, and service conditions can change the end-to-end wait.
For a production estimate, run the exact portrait and preset, then use the completed task details as the source of record. The useful comparison point is per completed output, not only the generation stage. Standard mode was used here because it keeps this test consistent; no unsupported claim is made about Pro mode being faster, better, or more expensive.
Which preset to pick
| Preset | Use it when | Caption that fits |
|---|---|---|
| Woman Birthday Mode | The post celebrates a woman and needs a fast birthday visual. | An age, first name, or short greeting. |
| Man Birthday Mode | The post celebrates a man with the same birthday format. | An age, first name, or short greeting. |
| Woman Zodiac Season | The content centers on a zodiac identity or seasonal horoscope. | One sign, such as ARIES. |
| Man Zodiac Season | The same zodiac format uses a matching man preset. | One sign, such as LEO. |
Choose the birthday presets when the occasion is explicit. Choose zodiac presets when the sign is the message. In both cases, start with a clean, front-facing portrait and resist the temptation to put a full sentence in the caption. For related tests, see AvatarMotion Multi: 6 Two-Photo Animations Tested, Live Avatar: Audio-Driven Talking Head Videos in 6 Tests, and Top 5 Viral Photo Effects APIs in 2026: 1 Selfie Test.
Open AvatarMotion with Caption on Wiro to choose a matching preset and test it with the actual portrait and caption intended for publication.
