왜 앞서 해보기를 진행하나요?

“Megan AI is an unusual blend of local LLM sandbox and real-time MetaHuman avatar. We need a wide variety of hardware, models, and use-cases to finish polishing stability, voice flow, and UX. Shipping in Early Access lets curious users try the tech now and steer our priorities with real-world feedback instead of guesses.”

얼마나 오래 앞서 해보기를 진행할 계획인가요?

“Our current roadmap points to 6–12 months in Early Access.
If big breakthroughs—or big surprises—occur, that window may shift. We’ll keep the community informed as we hit each milestone.”

정식 버전은 앞서 해보기 버전과 어떻게 달라지나요?

“The 1.0 release is planned to include:

Additional MetaHuman presets and environment maps

Expanded voice library and on-device TTS options

Long-term memory & summarised chat history

One-click model hot-swap and auto-download from Hugging Face

VR interaction prototype (head-tracked eye contact, gesture control)

Stability, performance, and crash-guard improvements
Scope may evolve based on community feedback; features will roll out as they’re ready rather than all at once.”

앞서 해보기 버전은 현재 어떤 상태인가요?

“Fully playable and fully offline.

Text or voice chat with a MetaHuman powered by any GGUF model you drop in

Three Apache-2.0 7 B starter models bundled

Real-time lip-sync, prompt editor, context window display, basic crash recovery

Settings menu for CPU/GPU layers, temperature, speech mode
Known limitations: occasional avatar animation pops, no long-term memory yet, UI is programmer-art, and some models may ignore safety prompts.”

앞서 해보기 기간 전후로 게임 가격이 달라지나요?

“Yes. Early Access is priced to reward early adopters. We expect to increase the price gradually as major features (memory, new avatars, VR mode) are shipped.”

개발 과정에서 어떻게 커뮤니티와 소통할 계획인가요?

“Steam Community Hub – fortnightly dev-logs and public bug tracker threads

Discord server for quicker feedback, model swaps, and scheduled beta branches

Opt-in experimental branch on Steam for users who want to test fixes before they hit the main build
We read every post, surface common issues in the Trello roadmap, and vote on feature priorities with the community.”