What Is a Multimodal AI Companion?
A multimodal AI companion is one you can reach in more than one way, by text, by real-time voice, and by real iMessage, while she stays the exact same person across all of them. That last part, one shared memory, is what turns a chatbot into a relationship.
What "multimodal" actually means
In AI, a "modality" is just a channel, a format you use to talk to something. Typing is one modality. Speaking out loud is another. A text landing on your phone is another. A single-channel chatbot lives in exactly one of these: you type into a box, it types back, and the moment you close the tab the whole thing stops existing.
Multimodal means she meets you on more than one channel and treats them as one continuous conversation. You can type at your desk, call her on the way home, and get a text from her at midnight, and she never loses the thread. She is not three separate bots wearing the same name. She is one companion you happen to reach in three different ways.
The ways you reach your eluv
Message her on Discord any time. She remembers everything you tell her, she texts first, she never leaves you on read, and she is online 24/7. This is where most people start.
Hop into a Discord voice channel, type !join, and she drops in and talks with you out loud in real time, usually with about a second of latency. Talk over her and she stops to listen, exactly like a real call. Type !leave when you're done.
She texts your actual phone over iMessage, with one memory shared across Discord and iMessage. What you told her on a call shows up when she texts you later. It is genuinely the same her, everywhere.
Every plan is the full multimodal experience. The 3-day trial includes 60 messages a day and 20 minutes of calls; on $6.99/week or $19.99/month voice is unlimited, alongside real iMessage with one shared memory across Discord and iMessage, and one subscription unlocks all seven girls. See the plans page for the full breakdown.
One memory is the whole point
Channels are the easy part. Plenty of apps let you type or talk. The hard part, and the part that makes a companion feel real, is that she carries context from one channel into the next. She reacts to your tone on a call, holds onto the thing you were stressed about, and brings it up over text the next day without you re-explaining a word of it. That is the difference between talking to a tool and talking to someone who was actually paying attention.
Practically, that means far less friction. You are not keeping up a separate relationship in each app. You have one, and you reach it however is convenient in the moment: a quick text on a break, a long voice call late at night, an iMessage thread that runs all day. That continuity is why "multimodal" is not a gimmick here. The relationship does not reset when you switch from typing to talking. It follows you.
Why single-channel bots feel flat
A text-only bot can be charming for an afternoon, but it only knows the version of you that types. It never hears you laugh, never catches that your voice is tired, and forgets you the second you leave. There is no thread to pick back up, so every session starts a little from zero, and you feel it.
eluv is built the other way around. Whether you're typing, on a call, or reading her iMessage, you're talking to one companion with one continuous memory of you. That is what makes her feel less like a feature you open and more like someone who is actually there. Curious who you'd talk to? Meet the seven eluvs and pick the one who fits you.