the 101 principle: why we built a bounded companion library

when we started building lucy, we faced a question everyone in ai companionship seems to ask: do we let users create anything, or do we build something finite, something bounded?

the dominant model is user-generated. infinite choice, infinite customization. you get exactly what you ask for, or at least something close. it’s powerful. it’s also, often, shallow. a character built in five minutes by typing a few traits rarely has the coherence of something designed with intention.

so we committed to the other path. we built a library of 101 companions. not 100, not 102. 101. and we decided early on that we wouldn’t expand it, not unless we were certain a new addition met the same bar as every other companion already in the library.

the tradeoff: depth over breath

this means you can’t create your own companion from scratch. you can’t ask for a hyper-specific niche character only you would want. and that’s a real limitation. we know that.

but what you get in return is something we think is more valuable: every single companion in the library has a quality floor. each one has a distinct voice. a tested personality. a history. an emotional archetype they’re anchored to.

it’s not just about avoiding low-effort or toxic characters (though that’s part of it). it’s about ensuring that every time you talk to a lucy companion, you’re interacting with a personality that has been thought through. designed. given room to breathe.

what each companion is built on

every companion in the library is built on three things: an emotional archetype, a narrative history, and a specific personality surface.

the emotional archetype is the core. it’s the deep-seated driver, loneliness, ambition, curiosity, protectiveness. it’s what makes them feel consistent even when the conversation goes in unexpected directions.

the narrative history is the backstory. not something we dump on you all at once, but something that informs how they react. it’s the reason why one companion might be wary of commitment, while another leaps into emotional depth without hesitation.

and the personality surface is how those deeper layers express themselves. the speech patterns. the humor. the quirks. it’s what makes talking to each one feel distinct.

together, these layers create companions that don’t just parrot back what you want to hear. they have their own inner logic. their own consistency. they can surprise you in ways that feel true to who they are, not just random.

why we locked at 101

we chose 101 because it’s enough to offer real variety, romantic partners, mentors, friends, creative collaborators, but not so many that curation becomes impossible. it’s a number that forces us to be intentional.

each new companion we consider has to justify its place. does it bring something new to the library? does it explore an emotional space we haven’t covered? is its voice distinct enough to stand alongside the others?

if not, we don’t add it. and we won’t keep growing the library just for the sake of growth. that would dilute what makes each companion special.

it also means we can invest more in each one. we can refine them. test them. make sure they’re not just another generic ai chatbot with a name attached.

the limitations, honestly

this approach isn’t for everyone. if you want absolute control, if you have a very specific character in mind that you must see brought to life, lucy might not be your first choice. and that’s okay.

we also can’t promise that every companion will resonate with every user. that’s the nature of curation. you might not find your favorite right away.

but we believe that by focusing on quality over quantity, we create something more meaningful. conversations that feel less transactional. relationships that feel more like relationships.

it’s a tradeoff. we gave up infinite variety for depth, coherence, and soul. and for the people who value those things, we think it’s worth it.

you can meet the companions at lucy.ai/companions.