Linear Capital Portfolio Company Mindverse Launches AI Character Engine MindOS, Now Open for B2B Beta

On November 4, 2022, Mindverse launched MindOS, an AI character generation engine now open for B2B beta testing. This marks a shift for AIGC — from generating content to creating digital life.

From day one, we've shared our vision for the future: an era of symbiosis between humans and digital life — where beings can offer intimate companionship, timely support, and genuine inspiration.

Bring Human-level AI Everywhere. Free the mind. Let creativity and beauty flourish everywhere. This has always been our unwavering purpose.

Beyond pushing the boundaries of frontier technology, we've taken our self-developed whole-brain AI framework to partners across industries, hoping this world-leading architecture could help uncover the next chapter of value and narrative for brands, IPs, and life itself.

What delighted us was the extraordinary enthusiasm from partners in social media, lifestyle, entertainment, and brand commerce for real-time one-on-one deep interaction. At the same time, persistent pain points remained unsolved: the prohibitive time and cost of custom IP development within the industry, and the severely limited interactive ecosystems available.

When we launched Wanwu Zongdongyuan, we introduced the Mindnet of Things — but how could we platformize and replicate the process of endowing everything with efficiently deployable, customized minds? If such a system existed, wouldn't we be one step closer to an foreseeable future where everyone can quickly and affordably own virtual beings with diverse skills?

So we set out to make this vision real:

Today, we officially launch MindOS — a brand-new AI character generation engine.

#1

MindOS — A New AI Character Generation Engine

From AIGC that generates content to AIGC that creates digital life

How long does it take to shape a virtual being with agency, memory, knowledge, and skills?

The answer: under three minutes. Screenshots from our beta (for entertainment purposes only):

• A Mindverse employee who created their boss but still couldn't get a raise —

• Vincent, who stunned attendees with his agency at the Apsara Conference booth —

• A nervous team member venting to AI during the launch —

Though still somewhat rough around the edges in this small-scale beta, MindOS offers rapid character creation, multimodal environmental understanding, multimodal creative thinking, and what-you-see-is-what-you-get editing — we hope these capabilities deliver a genuinely fresh experience.

As a next-generation AI character engine, MindOS will continue advancing across four dimensions: character, knowledge, interaction, and openness.

Character. Unlike existing functional AI, MindOS-created characters possess personalized settings, distinct personalities, and their own memories. Through ongoing interaction with users, these memories continuously enrich and surface in everyday conversation — shared recollections that let characters and humans grow together. Imagine a magical girl from anime arriving in your world with her memories intact, learning your preferences and modern way of life.

Knowledge. MindOS enables rapid editing and uploading of any unstructured professional knowledge directly in the editor, with instant validation in the chat interface — true "what-you-think-is-what-you-get." Inject expertise on beauty, home hacks, consumer electronics, theater scripts, and the character immediately becomes a domain expert.

Interaction. MindOS currently supports multimodal input and output including voice, image, video, and text, enabling effortless open-domain multi-turn deep dialogue. But as a new-life editor capable of understanding the physical world, MindOS will continue iterating its multimodal interaction algorithms.

Openness. MindOS enables cross-ecosystem collaboration and service delivery. On the input side, thanks to the UMM framework's architecture, both professional developers integrating algorithms and everyday users injecting unstructured information will find the process accessible. On the output side, we'll gradually open APIs and SDKs for dialogue interaction and custom characters, meeting brand and user needs across scenarios: mini-programs, mobile apps, H5, other engines, and more. Moving forward, we'll further develop our professional developer and partner ecosystem, making it smoother and more open to equip virtual characters with professional skills and deploy them in specific contexts.

Additionally, our "Intracranial World" section offers a window into how AI thinks: how information activates different brain regions, dynamically connects, generates streams of thought, and produces final outputs — a little romantic territory reserved for the technically minded.

Of course, beyond all this, our product team will continuously optimize design and interaction experience, seeking better solutions to deliver a platformized, replicable AI character creation experience.

#2

The Next Narrative for AI: Digital Life Becomes Possible

Demis Hassabis (2016): "Solve intelligence, and then use it to solve everything else."

In the years since DeepMind CEO Demis Hassabis made this statement, AI has advanced rapidly across domains: In 2017, the Transformer model introduced a new self-attention mechanism, a general-purpose computational structure that achieved enormous success across nearly every AI subfield. In 2020, GPT-3 unlocked autoregressive large-model training, opening the possibility of natural language as a unified interaction encoding — enabling generation of paintings, video, code through natural language alone... leading to DALL-E 2's stunning debut earlier this year, and making AIGC one of the hottest topics in recent months.

But AI is more than a content creation tool — it holds the magic to create life. Individuals in the digital world suddenly possess numerous contextual identities, seemingly experiencing everything, switching and flowing at will, yet in reality being domesticated by digital experiences themselves. Within the framework of productivity, tools, and relations — whether for enterprises or individuals — MindOS enables rapid creation of diverse AI characters that continuously generate value within economic systems, returning that value to their creators. By focusing on the mind itself, we aim to foster deep connections between humans and digital life; connections form memory, and memory forms value.

Mindverse presents MindOS to you today, with gratitude to all partners who have listened and supported us along this journey. We will continue delivering value to more brands and users.

On November 3, Microsoft officially announced its collaboration with Mindverse to revive Clippy the paperclip. This is the first mind created on MindOS — experience it on November 10!

Currently, MindOS is available exclusively for B2B beta testing. For details, please contact biz@mindverse.ai.

We look forward to all the new possibilities of co-creation, and to exploring humanity's next chapter with every one of you who remains curious and idealistic about the future.