Shot Supervisor

User authentication!
Create a project!
Utilize AI to generate images based on the project description in a story timeline
Keep all your shots in one place per project!
Use BRIA's API to make changes to image in seconds!
HDR control over images!
Export capabilities!

Inspiration

Shot Supervisor was inspired by a real gap in professional visual production workflows: the disconnect between creative intent and technical execution. Today, creators often jump between tools, prompts, exports, and manual tweaks just to keep shots consistent across a project. Prompt-based generation is powerful, but unpredictable, and it breaks down quickly when precision, repeatability, or professional standards like HDR, bit depth, and color space matter.

Bria FIBO’s JSON-native, deterministic control felt like the missing piece. Instead of fighting prompts, I wanted to build a director-style mobile tool where shots are planned, generated, edited, and exported with production-grade control—anywhere, on a phone.

What it does

Shot Supervisor is a mobile production companion for planning, generating, editing, and exporting shots with AI—designed for professional creative workflows.

The app is organized into three core tabs:

Projects

Create a project with a name and description.
Optionally generate up to 5 AI-generated shots automatically using an LLM-driven breakdown of the project description.
Generated shots are structured to maintain consistency across a project.

Shots

Each project contains multiple shots.
Create shots with a title, description, and generation prompt.
Generate images using Bria FIBO.
Attach a reference image from the camera roll or take a photo directly in-app to guide generation.
Regenerate shots while preserving intent and structure.

Editor

Edit any generated shot using professional-grade tools:
- Background removal
- Background replacement via prompt
- Blur background
- Quality enhancement
Advanced edit toggles:
- HDR controls (custom endpoint): exposure, contrast, highlights, shadows
- LUTs: None, Cinema (Filmic), Teal & Orange, Vibrant
- Color Spaces: sRGB, AdobeRGB, DCI-P3, Rec2020
- Bit Depth: 8, 10, 12, 16-bit
- Product editing via Bria Product API
Export final assets:
- Save to device or email delivery
- Formats: JPEG, PNG, TIFF, EXR

This turns a phone into a portable shot planning and post-production tool, not just a generator.

How I built it

Bria FIBO APIs
- Image generation
- Editing APIs
- Product APIs
JSON-native controllability
- Structured parameters for predictable, repeatable outputs
- HDR, bit depth, and color space controls highlight FIBO’s pro-grade strengths
Custom Vercel-hosted endpoints (for security and flexibility):
- bria-generate
- bria-edit
- bria-product
- export-hdr
- hdr-tone-map
- email
- llm-breakdown (uses DeepSeek-V3 to convert project descriptions into structured shot plans with bria's generate API)
Supabase
- Authentication
- Project and shot database
Mobile-first UX
- Designed for speed, iteration, and on-set usage
- Camera integration for real-world reference images

By separating generation, editing, and HDR processing into modular endpoints, the system is scalable, secure, and production-ready.

Challenges I ran into

Designing a professional editing workflow on mobile without overwhelming the user.
Balancing creative freedom with structured, deterministic control.
Integrating HDR and high bit-depth workflows in a way that feels intuitive rather than technical.
Building custom HDR endpoints to extend beyond default API constraints while remaining compatible with Bria’s pipeline.
Time constraints of a hackathon—some components were started but not fully polished.

Accomplishments that I'm proud of

Built a true professional-grade mobile tool, not just a demo.
Showcased Bria FIBO’s strengths:
- JSON-native control
- High bit depth
- HDR workflows
- Deterministic, repeatable generation
Combined AI planning (LLM breakdown) with structured image generation.
Created a scalable, secure backend architecture suitable for enterprise use.
Delivered an end-to-end workflow: concept → generation → edit → export.

What I learned

Professionals don’t want better prompts—they want better controls.
FIBO’s structured generation unlocks workflows that simply aren’t possible with prompt-only models.
HDR, color space, and bit depth matter deeply in real production environments.
Mobile can be a serious creative platform when paired with the right abstractions.
JSON-native generation pairs extremely well with agentic and automated pipelines.