ñawi kanki Tunche

• 

  nawiKankiTunche_poster_01

• 

  nawiKankiTunche_poster_02

• 

  nawiKankiTunche_poster_03

Inspiration

I’ve always loved Peruvian folklore, and bringing those stories into my work as a filmmaker has been a long-time dream. The found-footage format felt especially relevant today—our world is dominated by selfie videos and constant recording—so exploring that through AI was both challenging and exciting. I wanted the audience to feel fully immersed, as if they were experiencing the nightmare through Nayra’s eyes.

What it does

The short film exposes how greed can cost us everything we care about. It places the viewer inside a hostile jungle that feels alive, hunting, and inescapable. What sets it apart is the fusion of Peruvian folklore with a raw, intimate selfie-mode perspective rarely used in horror cinema.

How we built it

We created the film using a hybrid workflow with Hailuo, Veo 3.1, ChatGPT, and Gemini. Each long shot required 5 to 8 iterations to achieve stability, emotion, and movement. The most demanding technical challenge was maintaining continuity between fragments: starting with an initial frame, refining the clip until it felt natural, then using the last frame as the first frame of the next sequence to preserve seamless flow.

Challenges we ran into

Our biggest challenge was, again, continuity—keeping consistent acting, lighting, rain, and camera behavior. Some AI models drifted in style, refused to regenerate certain shots, or triggered safety filters, forcing rerouting of the creative process. And of course, as in any filmmaking effort, time and technical limitations were always present.

Accomplishments that we're proud of

We’re proud simply to have completed and shared the film—to demonstrate that cinematic storytelling with AI is not only possible, but creatively powerful. One of the moments that surprised us the most was the shot of the Tunche walking toward the camera; the tension and presence it achieved exceeded expectations. Overall, the project pushed us to understand AI limits more deeply.

What we learned

We learned that AI storytelling requires constant iteration, flexibility, and sometimes abandoning an approach to find a new path. Directing emotional nuance and tension through generative models is difficult and often unpredictable. And while we already knew Amazonian mythology well, the process reinforced how rich and haunting these oral traditions can be when translated into cinematic form.

What's next for ñawi kanki Tunche

There are no plans to expand this specific story, but we will continue exploring new short films with different themes to evolve our AI-driven filmmaking approach. The film will be submitted to festivals and shared across platforms to showcase what emerging tools can achieve in narrative cinema.

Built With

• chatgpt
• davinciresolve
• elevenlabs
• gemini
• hailuo
• veo