Estsoft co kr

Estsoft (이스트소프트, website https://estsoft.co.kr/ ) is an AI technology company positioned as "Beyond AI, we build the foundational interface for the AI world," providing solutions around conversational interfaces, content creation, multilingual communication, and digital avatars.

I. Main Products and Services

1. Perso Interactive (Conversational AI Avatar): A real-time conversational AI avatar service that can recognize and speak more than 100 languages, including English, Korean, and Chinese, and supports real-time conversation and translation. Its features include: generating more natural movements, expressions, and speech in an on-Device environment; understanding speech and expressions through Vision-Language and responding in real time; and expanding from a fixed location to an interface with mobility and immediacy.
2. Perso Interactive AI Station: An AI avatar kiosk for scenarios such as exhibition halls, hotels, airports, sports events, and public institutions. The official website lists uses such as exhibition guidance, product introductions, demonstration explanations, multilingual guidance, hotel digital concierge, event interaction, and library intelligent guidance, and emphasizes that it can operate 24/7, reduce manual burden, and automatically collect leads.
3. Perso Interactive SDK: An SDK for next-generation universal interfaces and interactions, usable from social media ads to high-performance websites, providing real-time conversation, support for more than 100 languages, and support for integrating various LLM model APIs such as OpenAI GPT-3.5 and HyperCLOVA X.
4. Perso Dubbing: An AI dubbing, video translation, speech translation, and lip-sync tool that supports dubbing in 32 languages and provides near-native natural voices and lip sync.
5. STUDIO PERSO: A professional AI avatar video production service that requires no filming crew, voice actors, or professional equipment, and can generate 4K high-definition videos in a few clicks; expert-level speech can be generated by selecting from more than 6,000 languages and voices; it supports uploading PPTs to automatically organize videos by scene; it supports uploading TXT, HWP, and DOCX documents as scripts; it supports free template modification, adding background images/videos/music, hiding the avatar in specific scenes, and one-click automatic subtitle generation and download. The official website states that videos generated after subscribing to a paid plan can be used for commercial purposes.
6. Other capabilities: The official website mentions AI that can understand Korean tone (automatically completing slides from content composition to reference materials), AI lip sync for content localization, and more.

II. Core Technologies

• Conversational AI Avatar: AIGenerative AI technology that fuses virtual humans with LLMs, supporting LLM model connection, appearance customization, voice customization, and personality customization.
• STV (Speech To Video): Input speech and video, modify the video by matching lip movements to the speech, involving Landmark detection and facial recognition.
• TTS (Text To Speech): Voice generation AI that converts text to speech, supporting IP customization.
• Speech Translation: Analyze the input video audio and translate it into the target language, while also analyzing the speaker's voice characteristics to make the translated audio close to the original speaker's timbre.
• AI Persona Creation: Generate an AI Persona through facial learning, endowing it with facial details, voice, expressions, and other styles.
• Multi-Agent System: Multiple specialized agents collaborate to find the optimal solution, adopting a multi-LLM architecture adapted to agents, involving planning support, tool-call optimization, cultural context understanding, multilingual search, and fact reasoning based on multiple references.

III. Applicable Scenarios The industries listed on the official website include: exhibitions and pavilions, airports and shopping malls, hotels and tourism, sports events, public institutions and libraries, educational institutions and course production, global enterprise internal communication and customer content, brand-localized product videos and customer testimonials, multilingual campaigns for marketing agencies, international distribution localization for media and entertainment, and global dissemination for content creators (YouTube, TikTok, Instagram).

IV. Partnerships and Projects The official website states that it has successfully executed more than 20 projects, with partners covering the public and education sectors, global enterprises, and others. The news section mentions the signing of an MOU with K Monster Group for the global business expansion of "Perso AI," the hosting of the EST AI Challengers 180-person hackathon, and reports related to the Perso Interactive business division.

V. Limitations and Information Not Provided The provided materials do not include specific pricing, differences between product versions, details of free trial quotas, SLAs, or a platform compatibility list. Implementation must be subject to the official page and business communication.