Spectrum by Prismatic

Social media buries intentional art under slop. Spectrum gives intentional creators the strategy to win anyway because depth always beats volume when someone finally shows it how to rise up.

Comment

Spectrum by Prismatic

What Inspired Us

We are musicians, illustrators, photographers. We have watched intentional creative work get buried under automated content that took four seconds to generate. We talked to artists at this hackathon who felt the same thing. The frustration is universal: platforms are not broken, they are working exactly as designed. AI content drives engagement, engagement drives ad revenue, and nobody is structurally incentivized to fix it.

Everyone building in this space is asking "is this real?" We decided that was the wrong question. The right question is: how do we make intentional art impossible to ignore?

What We Built

Spectrum is a strategy dashboard for intentional visual artists. It computes a Spectrum Score weighted heavily toward saves (4x multiplier) because saves signal real intent, not passive scrolling. The gold standard benchmark is a 3% save rate: saves / reach * 100.

The score is built around three signals:

  • Saves over likes -- saves mean someone wanted to come back
  • Process over product -- process content consistently outperforms final pieces with audiences that actually care
  • Depth over breadth -- we want the 100 people who will buy a print, not the 10,000 who scroll past

The weekly strategy brief reads like a creative director studied your work, not like an if-statement. Grounded, specific, yours.

The technical differentiator is Style Intelligence. Artists upload a five to ten second process clip or a finished piece. Spectrum sends it to a multimodal vision model (Groq + Llama 3.2 Vision) with a structured prompt that extracts visual style elements: brushwork, palette character, composition tendencies, medium type, texture quality. The model returns a plain language style profile that gets folded into every recommendation.

Not: "Post more process content."

But: "Your brushwork is your signature. Each bristle is visible and unmistakably made. Your three highest-save posts all lead with that texture. Your next Reel should open on a three-second close-up before pulling back. Artists in your niche who lead with texture see 2.3x higher save rates."

How We Built It

  • Frontend and backend: Lovable
  • Auth and database: Supabase
  • Style analysis: Groq + Llama 3.2 Vision
  • Consistency framework: Spectrum Rubric prompt document
  • Scoring logic: Save-weighted engagement algorithm with 4x save multiplier and 3% benchmark

What We Learned

Slop is not an AI problem. Slop is an intention problem. AI is not always slop. Slop is low effort and no intention. The process is the differentiator, not the tools.

We also learned that trying to detect AI content is a losing battle technically and economically. Platforms will never fix this because AI content drives the engagement that drives their revenue. Detection tools are in an arms race they cannot win. The only durable solution is to make intentional work so strategically sharp that it finds the audience already looking for it.

Depth is human. That is the whole philosophy.

The Math

The Spectrum Score weights engagement signals as follows. Given likes $L$, saves $S$, comments $C$, shares $Sh$, and reach $R$:

$$\text{Spectrum Score} = \frac{L + 4S + 2C + 2Sh}{R} \times 100$$

The save rate benchmark:

$$\text{Save Rate} = \frac{S}{R} \times 100 \geq 3\%$$

Challenges

The hardest problem was not technical. It was philosophical. We spent hours killing approaches that did not work: detection tools, walled gardens, authentication claims. Every path that tried to prove humanity was either technically broken or economically backward.

The breakthrough was reframing entirely. We are not a detection tool. We never claimed the process upload proves anything about who made the art. It helps us give better advice. That framing is fully defensible and it is the right one.

The second challenge was making the strategy brief actually sound like a creative director and not a conditional statement. The difference between "your save rate is below 3%" and "your brushwork is your signature and your audience already knows it" is everything. That is the product.

What Is Next

The Spectrum Score and Style Intelligence work for any visual medium. Digital illustration is the beachhead. Photography, traditional art, and music content creators are next. No new infrastructure required -- the niche selector in onboarding expands and the style analysis prompt adapts to the medium. The core product loop is identical.

1% penetration of the monetized TAM = $12.96M ARR

Built With

Share this project:

Updates