We will be undergoing planned maintenance on January 16th, 2026 at 1:00pm UTC. Please make sure to save your work.

Inksfire

Ink's on fire man!

Comment

Inspiration

The inspiration for this project came naturally. I’ve always loved watching movies—but more than that, I’m drawn to the process behind them. Film is one of the most powerful mediums for storytelling, and I’ve long wanted to merge the two passions that shape me most: software and stories.

Stories play an essential role in our lives—we create them every day. In fact, reality often proves stranger than fiction, and imagination knows no limits. Even in the midst of the digital revolution, many still dream of being part of worlds like the Hogwarts universe.

This project is my first step toward building a future company. Inksfire is designed to harness the power of data and AI—not to replace creativity, but to challenge and elevate it, helping creators refine their own unique taste and voice.

What it does

Inksfire helps users write stories—freely and with structure. Writers can simply write, or opt to enhance their ideas with data-backed insights.

The platform connects with Qloo and Perplexity to suggest tags that describe a story’s themes and concepts. Each tag comes with demographic data—enabling writers to understand their story’s potential audience. After selecting tags, an aggregate demographic profile is created. When a story is submitted, Gemini LLM evaluates whether the selected tags are actually reflected in the content, producing a “projected demographic” from what was met or missed.

The system also generates a character list, character relationship map, and story beats. Writers can describe a character and prompt for actor suggestions—Perplexity finds possible matches, and Qloo returns audience data for those actors.

There’s also a metadata comparison tool that allows users to analyze two different movies based on tags and demographics.

How we built it

The app is built using Angular and Firebase for the frontend and backend infrastructure. Gemini API powers the natural language understanding and evaluation. Qloo provides tag-based cultural and demographic insights. Perplexity is used for search and real-time data retrieval across domains. Together, these tools orchestrate a workflow that turns ideas into structured, analyzable story content.

Challenges we ran into

One major challenge was understanding and applying unfamiliar data—especially around demographic analysis. Working with layered APIs like Qloo and Perplexity added complexity to the orchestration.

Additionally, part of the data model for comparing entities was changed midway through the hackathon, which disrupted the logic for comparing movie metadata. That setback affected the quality of the movie analysis component during the hackathon.

Another challenge was scoping: this project represents a fraction of a much larger vision, and fitting that into a working prototype within time constraints required tough prioritization.

Accomplishments that we're proud of

This is the first working prototype of Inksfire—designed, built, and integrated solo. While there are still loose ends to tie up, it brings together AI, real-time data, and creativity in a meaningful way. Seeing ideas take shape and watching a system help evaluate and structure stories was a major personal milestone. Seeing the top of the ice berg makes me happy.

What we learned

This was my first experience working deeply with external datasets and understanding how to use data, not just display it. While the execution wasn’t perfect, I pushed my own limits in exploring and applying data to augment creativity. I also learned the value of structured storytelling tools, and how AI can be a creative collaborator, not just an assistant.

What's next for Inksfire

Fix UI bugs and polish the feedback/evaluation loop Improve overall usability by resolving visual glitches and ensuring that the AI-generated feedback on tags and demographics is timely, accurate, and easy to interpret.

Add screenplay writing tools linked to stories Introduce a dedicated script editor that automatically structures content into scenes, dialogues, and stage directions based on the original story.

Use tags to generate automatic scene breakdowns and flows Leverage selected story tags to auto-generate scene-by-scene outlines, suggesting pacing, transitions, and thematic progression.

Build mock 3D scenes for visual ideation Provide a simple 3D environment where users can block out scenes, experiment with camera angles, and visualize key moments.

Generate rough AI sketches for scene shots Use AI image generation models to produce concept art or frame thumbnails for each major scene, helping creators with visual storytelling.

Add mock table reads using ElevenLabs for voiceovers Enable users to hear their scripts read aloud using AI-generated voices, simulating actor table reads for character tone and rhythm evaluation.

Built With

Share this project:

Updates