KRAFTON launched the Kiff K-pop AI chat service in South Korea, according to an Inven Global report. A K-pop AI chat service is software allowing fans to exchange bilateral messages with virtual idol personas. While services like Kiff focus on text threads, companion platforms such as Prelulu provide real-time face-to-face video calls where characters speak in their own voice, reflect emotion in their facial delivery, and support mid-sentence interruption.
KRAFTON's Kiff Launch and What a K-Pop AI Chat Service Is
Video game publisher KRAFTON launched a dedicated K-pop AI chat service called Kiff in South Korea, as reported by Inven Global. A K-pop AI chat service provides direct conversational interaction with synthetic idol personas rather than broadcasting static agency updates. The Inven Global report confirms the South Korean debut of Kiff as an AI chat offering, though the report does not provide pricing schedules, system latency figures, or mention live video capabilities.
Traditional celebrity fan messaging applications rely primarily on one-to-many subscription broadcasts, where human artists post predetermined text updates or media clips to broad subscriber channels. By contrast, an AI chat service uses natural language systems to generate bilateral, context-specific responses to each incoming fan message. For companion platform users, KRAFTON's launch highlights an industry shift toward interactive fan dialogues, moving beyond one-sided feed consumption into direct conversational threads.
While branded services center on text exchanges, companion platforms such as Prelulu, an AI character-companion app with live video calls, offer real-time spoken and visual sessions. You can explore interactive conversational formats through an ai video chat to observe how direct vocal responses differ from text queues.
Live Video Calling Mechanics Versus Text-Based Idol Messaging
Live avatar calling replaces asynchronous text bubbles with a synchronized audiovisual conversation where the character speaks in their own voice. On Prelulu, an AI companion app with live video calls, you see and hear the character answer you face to face, with their face and delivery carrying emotion as they speak.
In a text-based setup, the exchange depends on simulated typing delays, message boxes, and occasional pre-rendered media attachments. Engaging in a live call introduces real-time spoken pacing. On Prelulu, you can cut in mid-sentence during a voice or video call to redirect the topic or add an observation, functioning like a standard telephone call.
User camera controls operate independently from the persona's video feed. During a call on Prelulu, your camera remains turned off unless you choose to turn it on, and sharing your camera requires explicit permission before the character can see what you present. You can review these privacy boundaries in our guide to Face to Face AI Chat.
Context Retention and Scenario Flow in Virtual Idol Conversations
Context retention allows an AI companion to carry details from earlier text discussions into active calling sessions. On Prelulu, an AI companion app with live video calls, saved relationship details and recent chat context can carry into voice and video conversations, though memory is selective rather than a complete verbatim transcript.
When roleplaying with idol personas, selective memory systems reference recent topics such as rehearsal notes, tour stops, or shared background lore without requiring you to restate your context at the start of every session. This contextual continuity bridges text chat with vocal calls across repeated sessions.
Catalog depth dictates the variety of conversational situations available. As of 2026-11-02, more than 100 characters in Prelulu's public catalog can be video called. Callers can choose between backstage preparation scenarios, creative collaborations, or unstructured off-duty conversations across the platform's roster.
Platform Architectures: Modalities, Access Models, and Call Economics
Evaluating conversational services requires examining supported interaction modes, documented camera controls, and credit structures. Prelulu publishes this analysis and develops the Prelulu platform. Prelulu is accessible on Web at prelulu.ai, iOS, and Android, and the prelulu.ai website interface is available in 27 languages as of 2026-11-02. The free plan includes a limited daily text allowance, while voice and video calls use credits.
Competitor models handle modalities and billing through distinct mechanisms. On Kindroid, direct web subscribers pay $13.99 for 1 month, checked 2026-09-02 on kindroid.ai/v2/docs/subscriptions/. Kindroid subscribers can access Live Avatar Video with real-time lip-sync and gestures using an avatar photo or optional Driving Image, checked 2026-09-19 on kindroid.ai/v2/docs/voice-calls-and-video-calls/. That documentation specifies that Standard Live Video costs 2,000 audio credits per minute, Premium Live Video costs 4,000 audio credits per minute, and callers can enable their camera so the AI can see it, with mobile camera video requiring the app to remain in the foreground.
Other platforms restrict visual inputs rather than generating an avatar video feed. On Replika, real-time video recognition is gated to the Platinum tier, which also allows Training Mode up to 100 times per week, Read Replika's Mind for up to 50 messages per week, and up to 10 realistic selfie videos, checked 2026-09-01 on help.replika.com. Replika's documentation checked 2026-09-19 clarifies that Platinum includes real-time video recognition via camera input, which must not be confused with generating an avatar video or a selfie clip, and that current subscription prices are displayed inside the app. For additional technical comparisons across platforms, consult our overview of AI Character Video Call Online.
Where the Short Answer Stops Being True
The premise that open companion platforms replace specialized K-pop AI chat services stops being true when official agency intellectual property, authorized music catalogs, and canonical idol schedules are required. Services like Kiff from KRAFTON cater to fans who want structured interactions tied specifically to official agency personas, approved promotional narratives, and localized fan club ecosystems in South Korea.
Open companion platforms offer original fictional characters and unscripted conversation, but they do not provide authentic agency-cleared updates or real-world concert ticketing access. When your primary objective is tracking the real-world activities of a specific licensed pop group, an official agency-aligned text tool fulfills that role in a way open roleplay platforms cannot.
The short answer also ceases to apply when credit economics favor asynchronous text over continuous audiovisual media. Because voice and video calls consume credits across platforms like Prelulu and Kindroid, long-form daily engagement remains far more sustainable in text messaging unless you budget specifically for per-minute audiovisual usage.
Meet your AI character face to face
Choose or create a character for live video chat, roleplay, shared stories or company. Check the available call credits and plans before starting.