Creating a Cinematic Short Inspired by Honduras and Mayan Mythology

"B’alam: El Amanecer del Guardián" (The Dawn of the Jaguar Guardian) is a short cinematic concept inspired by Honduran mythology, Mayan artistry, and the enduring symbolism of the jaguar guardian.

The story draws from the belief that when the sun disappeared each evening, it transformed into a jaguar and traveled through the underworld until dawn. I wanted to create a short visual concept that felt rooted in that mythology while exploring what AI video generation can do when paired with human editing and post-production decisions.

https://www.youtube.com/watch?v=bpHYB87MdgU

This project began as a creative challenge: build something powerful using only the three free Veo generations available to me. From ideation to export, the full concept took just four hours.

Step 1: Use AI to generate base clips

I used Veo through Google Gemini to generate the base visuals for this concept. The goal was not to get a perfect final result in one try. The goal was to generate strong raw material I could shape later through editing.

First Try: FAIL!

My first attempt asked for too much.

This was the exact prompt I gave Veo:

[00:00–00:02] Sweeping sunrise over Copán Ruinas. Mist curls around ancient Maya temples as tropical birds call. Golden light reveals glyphs etched in stone.

[00:02–00:04] Close-up: an explorer brushes moss from a jaguar statue. Its eyes glint. A deep rumble vibrates through the ruins.

[00:04–00:06] The statue awakens — glowing veins of gold race across its body. Jungle drums and flutes rise.

[00:06–00:08] Wide shot: the Jaguar God towers above the temple, half spirit, half stone, roaring as light and leaves swirl. Emotion: awe and reverence.

style: Pixar-style animation

lighting: golden hour cinematic

color palette: lush greens, golds, turquoise

tone: mystical adventure

environment: ancient Maya temple, rainforest canopy

camera: sweeping drone, cinematic focus

• Result on my Facebook Page: First video produced by Veo

• Takeaways: I needed to simplify the ask. With AI video, asking for fewer actions usually leads to a stronger result.

  
  

Second Try: Good Base

For the second attempt, I simplified the prompt and focused on getting a usable base clip instead of a perfect sequence.

This was the exact prompt I gave Veo:

[00:00–00:02] Sweeping sunrise over Copán Ruinas. Mist curls around ancient Maya temples as tropical birds call.

[00:02–00:04] Close-up: an explorer brushes moss from a jaguar statue. Its eyes glint. A deep rumble vibrates through the ruins.

[00:04–00:06] The statue awakens — glowing veins of gold race across its body. Jungle drums and flutes rise. The explorer is in awe.

[00:06–00:08] the statue runs towards the camera and opens it’s mouth as if it’s about to eat the viewer

style: Pixar-style animation

lighting: golden hour cinematic

color palette: lush greens, golds, turquoise

tone: mystical adventure

environment: ancient Maya temple, rainforest canopy

camera: sweeping drone, cinematic focus

• Result on my Facebook Page: Second video produced by Veo

• Takeaways: This version gave me something much more usable.

  
  

  The timing was still off. The explorer reacted before the jaguar’s eyes lit up, which weakened the moment. But the clip had a much stronger foundation, and that made it worth keeping.

  
  

  I also ended up reusing part of the opening scene later in the final short. After the Guardian transports us into the ancient Mayan world, I spliced part of this clip back in to support the transition.

  
  

  Another thing worth noting: in the generated version, the Guardian does not spit fire. That effect was added later in post-production.

  
  

Third and Final Try: Continuity

Continuity is one of the biggest challenges with AI-generated video.

To improve my chances of getting a better transition, I took a screenshot of the exact frame where I wanted the next clip to begin. Then I used that screenshot as the reference image for the final generation.

This was the exact prompt I gave Veo:

Using this image as your initial frame, create a scene where the jaguar swallows the viewer, the jaguar pounces, swallows the viewer the screen dips to black before revealing a vibrant Mayan City of Copan in Honduras when Mayans were alive and thriving . style: Pixar-style animation

lighting: golden hour cinematic

color palette: lush greens, golds, turquoise

tone: mystical adventure

environment: vibrant Maya temple, rainforest canopy

• Result on my Facebook Page: Third video produced by Veo

• Takeaways:  I needed to be more specific about color continuity. I should have told Veo to keep the colors exactly as they appeared in the reference image. Because I did not, the generated clip came out with a different color balance, and I had to color correct it in Premiere Pro so it would match the rest of the film.

  
  

Step 2: Post Production Magic

The AI generations gave me the raw material, but post-production is what made the concept feel like a finished piece.

I used Adobe Photoshop, Premiere Pro, and CapCut to refine what Veo generated and shape it into a more coherent short film.

What I fixed in post-production

1. TimingIn the second clip, the explorer reacted before the jaguar’s eyes lit up. I adjusted the sequence to improve the rhythm of the moment and make the reveal feel more intentional.

2. ContinuityI reused and spliced footage from earlier generations so the final concept felt more connected from shot to shot.

3. Color correctionThe third generation shifted away from the established palette, so I color corrected the footage in Premiere Pro to bring it back in line with the lush greens, golds, and turquoise tones I wanted throughout the film.

4. Added visual effectsThe Guardian did not spit fire in the original generated clip. I added that effect later in CapCut to heighten the mythical energy of the final sequence.

The real lesson

This project reminded me that AI does not replace creative direction. It speeds up ideation and gives you raw visual possibilities, but the final result still depends on editing decisions, continuity, pacing, and taste.

The prompt matters. The reference image matters. But what you do after generation matters just as much.