AI Travel Photo — AI skill for Claude Code

ai-travel-photo

A Claude Skill that puts **one specific person's face** into a real place, and returns something that reads as an actual travel photo — not an AI portrait standing next to a landmark.

Hội An at lantern hour · the Sa Pa rice terraces · Hạ Long Bay at sunrise · the Huế Imperial City · Paris at golden hour · Tokyo in cherry-blossom season.

*[Tiếng Việt bên dưới ↓](#vi)*

---

What makes it different from a plain image prompt

A prompt gives you one render. This skill runs the whole job:

	A prompt you write	This skill
The face	drifts into a different person on every render	a 7-group identity anchor, pasted verbatim into every render
The place	"Hội An" → a generic old-town image	18 destination recipes with a named vantage point, the hour that place is best, colour grade and wardrobe
You don't know where to shoot	it guesses	it researches, then offers 4–6 options spread across scene type, hour, distance and mood
Sign-off	you post it	a mandatory 7-point face check before delivery, capped at 3 attempts

Examples

The same face, four places. **The face in these images is fully AI-generated — it is not based on any real person's photo.**

**Step 3 — the skill offers a menu before it renders anything.** Given only a city name it does not jump to an image; it proposes four to six distinct treatments, each with the best hour, a pose, wardrobe and colour grade. *(These boards are in Vietnamese; the skill writes them in whatever language you use.)*



**Steps 5–7 — once you pick one, it renders and checks the face before handing it over:**

Đà Lạt, dawn mist	Paris, golden hour

Three more in [`Images/`](Images).

What you need

• An assistant that supports Skills.
• **A reference photo of the face — required, not opti