Synorigin

Checking a character concept for originality

A character built on a single, tightly-held premise - and what an originality check found underneath it.

This result is from a Claude Sonnet 5 session using the exact system prompt now in production, run before live web search was integrated. Reasoning and tone match the live app; matches were found from training knowledge, not live search.

What was being tested

Whether the check would recognise that Willow's nature is a structural, physical limitation rather than a personality trait or a curse - and whether the gap would read as a response to this specific text rather than a template applied to anything.

The idea submitted

Willow is shadow. Not haunted by shadow, not associated with it - she is it. And shadow, by nature, is the one thing that can never be sun. Her touch drains warmth rather than giving it. She reaches toward Marcus, and stops short.

What Synorigin returned

Closest matches

The originality gap

Every trait derives from the same unbroken rule. The logic is consistent, which is a strength, but nothing about her contradicts or complicates the root metaphor - so she reads as an extended metaphor rather than a person with an interior.

The pivot

Give her one deliberate choice that cuts against her own physics, rather than another trait that illustrates it again.

Verdict

Roughly 70% recognisable as written - the gap above is what closes it.

Why this result matters

The matches span three different media - comics, film, and a genre convention rather than a single title. That matters: a check that only names the most famous possible comparison is pattern-matching on fame, not on structure.

The gap is also not "make it more unique." It identifies a specific structural property - every trait reducing to one root metaphor - and gives an action that would change that property. That is the difference between criticism and a usable note.

Related

Check your own idea