Aethos

Aethos enriches basic image prompts with deep cultural context using the Qloo API, making AI-generated prompts feel locally authentic.

Comment

Inspiration

We realized that while AI-generated images are powerful, they often miss the cultural depth that makes a place feel real. We were inspired to build something that could take a basic idea and layer in rich, local context, something only someone living there might know. The goal was to make AI feel more human and culturally aware.

What it does

Aethos takes a simple prompt and enhances it with cultural elements based on a specific location. By using the Qloo Insights API, it pulls data like local places, brands and weaves them into a new, more grounded version of the original prompt. The user can then use this with any AI image model to get much richer visual results.

How we built it

We used the Qloo Insights API to extract culturally relevant attributes tied to a location. We mapped user input (a base idea + location) into a structured query that returned various cultural insights. We then used prompt-engineering techniques to blend these into a coherent visual prompt that could be plugged into tools like Imagen(most recommended) or Midjourney.

Challenges we ran into

While I didn’t integrate full image generation (thanks to the lack of free-tier access to SOTA models), the enhanced prompts work seamlessly with external tools like Imagen(most recommended) or Midjourney. Another challenge was figuring out which cultural dimensions from Qloo actually improved the final outputs. It took time to find the right combinations.

Accomplishments that we're proud of

We’re proud of how effective the final prompts turned out. Even simple inputs started to feel deeply personal and locally inspired once enriched. The system was also fast and clean, it gave users strong creative control without overwhelming them.

What we learned

We learned how powerful cultural metadata can be in generative workflows. Understanding and fine-tuning Qloo's API was essential, it’s not just about calling endpoints but knowing which kinds of insights translate best to visuals. We also gained a deeper appreciation for how small tweaks in prompt structure make a big difference in AI output quality.

What's next for Aethos

We want to integrate native image generation once access to models like Imagen becomes possible. We’re also looking to support more granular cultural layers—like historical time periods or niche subcultures. We see Aethos evolving into a creative tool for artists, storytellers, and marketers who want their AI visuals to feel truly situated in place.

Built With

Share this project:

Updates