All products

Blend poetry into photos with AI

Write a short original Chinese poem from your photo’s details, then integrate the text with the scene’s colours and surfaces while preserving the original subject.

Blend poetry into photos with AI by starting with the actual scene rather than a stock quotation. The photo-poem-fusion skill uses subject, light, colour and atmosphere to write a short Chinese poem, then plans lettering that belongs on a wall, glass, shadow or suitable part of the image.

It suits authorised personal or creative photos. The original workflow is for Chinese verse; equivalent results in another language or script are not established. Scene identity, subjects and the number of people should stay intact. Outputs are separate versions and need direct visual checks of text, dimensions and aspect ratio.

What it does

  • Write a poem from visible scene details
  • Choose lettering colours from the photo
  • Integrate text with an appropriate surface
  • Preserve subjects and scene identity
  • Compare distinct visual treatments
  • Inspect every character’s legibility
  • Check actual output dimensions and ratio

Install and use

  1. Install the skill in an image-capable agent with an authorised image-generation or editing tool. Accounts and credit depend on your host tool; use photographs you may process.
  2. Specify the photo and constraints: “Write a short Chinese poem for this scene and integrate it with the glass, keeping the subject, lighting and frame unchanged.”
  3. Review the poem and placement before rendering. Decide whether private details or people’s images may go to an external image service; identifying them is unnecessary.
  4. Open the finished image and check each character, placement and actual dimensions. Request a local text correction when only a small area is wrong, preserving the rest. Keep versions with incorrect text or proportions labelled as drafts.

FAQ

Source

Made by atian-create. Original skill: photo-poem-fusion, published on GitHub under the MIT licence.

View on GitHub