Have you ever seen a stunning photograph and thought, "I wish I could create AI art inspired by this"? With our Image to Text tool, you can do exactly that—and take it even further. In this tutorial, we'll show you how to extract a detailed prompt from any image and then transform it into entirely new creative directions.
Step 1: Start With a Real Photograph
For this example, we'll use a beautiful photograph of the Dervish monastery (tekke) in Blagaj, Bosnia and Herzegovina. This historic building is dramatically nestled against a cliff face with turquoise waters flowing beside it—a perfect subject to work with.
Our source image: the famous Blagaj Tekke in Bosnia and Herzegovina.
Step 2: Extract the Prompt
Head to AI Tools → Image to Text. Upload your image and select the first mode: "Image Prompt". This mode is specifically designed to analyze images and generate prompts suitable for AI image generation.
Select "Image Prompt" mode in the Image to Text tool.
Step 3: Review the Generated Prompt
After a few seconds, the AI generates a comprehensive prompt describing every detail of the image:
A picturesque, traditional Dervish monastery (tekke) in Blagaj, Bosnia and Herzegovina, nestled against a massive, textured cliff face, full shot. The monastery is a multi-story, white-washed building with a distinctive black and white curved roofline, black window frames, and several dark, rectangular windows, some with black shutters. A prominent white chimney rises from the roof. The building features an overhang on the upper floor extending over the water. To the left of the building, a calm, clear turquoise river flows, reflecting the surrounding light and colors. Lush, vibrant green foliage and trees frame the scene, with some leaves in the foreground appearing slightly out of focus, creating a sense of depth and natural framing. The cliff face behind the monastery is rugged and varied in texture, with shades of grey, brown, and hints of green moss and small plants growing in crevices. A large, dark cavern opening is visible in the upper right portion of the cliff. The lighting is bright and natural, indicating daylight, with clear shadows suggesting a sunny day. The overall mood is serene, ancient, and majestic, highlighting the harmony between architecture and nature. Realistic photography, high detail, natural colors.
Notice how detailed the prompt is—it captures the architecture, colors, lighting, composition, mood, and even the relationship between the building and its natural surroundings.
The AI extracts a detailed prompt from the photograph.
Step 4: Generate With the Prompt
Now take this prompt to the Generate tool and create AI images. We tested it with different models, and SeeDream produced the best results—capturing the essence of the original while adding its own artistic interpretation.
AI-generated result using the extracted prompt with SeeDream model.
Step 5: Get Creative—Edit the Prompt
Here's where it gets exciting. What if you love the composition but want something more exotic? Go back to the Image to Text tool and use the "Edit with AI" feature to transform your prompt.
We entered these instructions: "Rewrite this to be on the Moon."
Using "Edit with AI" to transform the prompt to a lunar setting.
Step 6: The Lunar Version
The AI rewrote the entire prompt, keeping the architectural concept but reimagining it for a lunar environment:
A futuristic lunar Dervish-style house, Tekke, built into the side of a towering, rugged crater wall, prominently featured in the upper-midground. The house is bright white with a dark, almost black, curved roof and dark metal trim, window frames, and shutters, contrasting sharply with the light walls. It has multiple stories with small rectangular windows and a distinctive projecting upper section. A single white antenna rises from the roof. The crater wall behind the house is a complex texture of grey and brown lunar regolith, with a dark, deep cave opening visible to the right of the house. In the foreground, sparse, translucent lunar flora with small, oval-shaped leaves partially obscures the view, creating a natural frame. To the left of the house and in the lower-left corner, a calm, clear turquoise-green liquid flows, reflecting the distant Earth and surrounding lunar landscape. The lighting is bright and natural, suggesting a sunny lunar day, with strong highlights on the white house and the translucent flora, and soft shadows defining the textures of the crater. The overall mood is serene, ancient, and picturesque. The style is realistic photography.
The monastery reimagined on the Moon—same composition, completely different world.
Step 7: Try Another Direction—Maldives
Let's try one more variation. This time we instructed: "Rewrite this to be on the Maldives, but keep the house inside the cave."
A picturesque, traditional Dervish monastery (tekke) in the Maldives, nestled within a massive, textured cave face. The tekke is a whitewashed building with a dark, wavy roof made of stone shingles and distinct black trim and window frames. It features multiple stories, dark rectangular windows, and a prominent chimney. One section of the building has a cantilevered bay window. To the left of the tekke, a calm, turquoise-green ocean flows, reflecting the surrounding environment. Lush, vibrant green foliage, including palm trees and tropical bushes, frames the scene. A large cluster of bright green leaves is prominently in the foreground on the right, partially obscuring the view of the ocean and the tekke, creating a sense of depth. The background is dominated by the steep, rough limestone cave, showing varied textures, shades of grey, brown, and some reddish hues, with patches of green moss and small plants growing on it. The tekke is visible within the cave entrance. The lighting is bright and natural, suggesting a clear day, with sunlight illuminating the tekke and the foliage, creating strong highlights and shadows. The overall mood is serene, historic, and natural.
The same concept transported to a tropical Maldives cave setting.
The Possibilities Are Endless
This workflow opens up incredible creative possibilities:
- Start from reality – Use any photograph as your foundation
- Extract the essence – Let AI capture what makes the image work
- Transform at will – Edit the prompt to explore new directions
- Generate variations – Create AI art inspired by real-world compositions
Whether you want to recreate a scene faithfully, transport it to another world, or use it as a starting point for something entirely new, the Image Describer tool makes it easy.
Bonus: From Images to Video
Want to take your AI-generated images even further? We took the images created from this tutorial and transformed them into a cinematic video using our AI Video Generator. Watch how the monastery comes to life across different worlds:
The monastery journey: from Bosnia to the Moon to the Maldives—all generated from a single photograph.
Comments (10)