Moemate was a decentralized AI ecosystem for creating and interacting with custom multimodal AI agents with chat, voice, memory, and 3D avatars, but its consumer app shut down in February 2025 and the original product is no longer operational.
Moemate combined large language models, computer vision, speech synthesis, and 3D rendering so users could build characters with their own personality, voice, and appearance, then later added a token layer around Mates.lol, MATES, MOE, and the flagship agent Nebula.
For AI researchers, developers, skill builders, agent creators, and users exploring advanced customizable AI agents, Moemate is still a useful case study in how decentralized and tokenized AI products were structured. Official sites such as moemate.io are no longer a live product destination, and while MATES and MOE may still trade on Solana, they are not tied to the original working platform. This article breaks down Moemate’s features, underlying tech stack, token design, ecosystem pieces such as Mates.lol and MatesDAO, how users created agents on the platform, and what the shutdown says about the opportunities and risks in democratized AI agent ecosystems.
Takeaways
Moemate was an AI-agent platform for custom characters with chat, voice, memory, LoRA, and 3D avatars.
The stack combined LLMs such as GPT-4 and Claude, computer vision, text-to-speech, and 3D rendering.
The ecosystem was built around Moemate.io, Mates.lol, MatesDAO, MATES, and MOE.
Nebula was the flagship AI character, linked to the $MOE token.
MATES launched on Solana in January 2025 with a total supply of 19.5 billion.
The consumer app shut down in February 2025 and is no longer operating.
MATES and MOE may still trade, but residual liquidity and utility are limited.
Token trading after shutdown should be treated as speculative, not as access to the original product.
What is Moemate?
Source: Moemate website
Moemate was a platform that created advanced AI agents with multimodal and multimedia capabilities. It used large language models, computer vision, speech synthesis, and 3D rendering to let users build AI characters that could talk, remember context, and take on specific tasks.
Users could design characters with unique personalities, voices, and appearances. The intended use cases ranged from gaming and entertainment to education and productivity. After the February 2025 shutdown, these features should be read as a description of the former product, not as a live walkthrough.
The main goals and missions of Moemate
The main goal of Moemate is to democratize artificial intelligence, making it accessible and customizable for various use cases. The platform was co-founded by Akhad Shams and Chris Larkin with the aim to empower users to create and interact with personalized AI agents.
Moemate Features
Talking to Characters: Moemate lets users interact with each ai companion in real time through dynamic personalized conversations and creative roleplay scenarios. These interactions can build a stronger emotional connection, making the experience more immersive and interactive.
Skills: AI characters on Moemate can be equipped with various skills to perform specific tasks. These skills range from simple commands to complex actions, and certain features also include multimodal tools like image generation and web searching.
Create Character: Users can use character creation tools to build unique AI characters with personalized attributes. This level of character customization covers appearance, personality, and abilities, and the platform also offered a large character library with over 4,000 available characters.
Voice cloning: Moemate provides voice cloning function, allowing users to clone voices for their AI characters. This feature adds realism to interactions, making conversations more vivid and interesting.
Customize LORA: Users can use customized LORA models to accomplish specific tasks. These models enable AI characters to perform specific functions based on user-defined parameters.
Knowledge Upload: Users can upload knowledge to their AI avatars to enhance their ability to provide accurate and relevant information. This feature enables AI agents to continuously learn and improve.
Long-term memory: The AI characters on Moemate had a long-term memory system that remembered previous conversations, allowing chats to continue seamlessly. This made interactions more meaningful by supporting deeper relationships over time.
Public Memory: Users can share memories with other users through the public memory function. This makes collaborative experiences possible and enables the sharing of valuable information and interactions.
The technology behind Moemate
Large Language Model
Moemate uses large language models like GPT-4 and Anthropic's Claude for text generation and understanding. These models are designed to process and generate human-like text, enabling AI agents to engage in meaningful and coherent conversations.
GPT-4: Developed by OpenAI, GPT-4 is an advanced language model known for generating high-quality text. It can understand context, respond to queries, and generate detailed and relevant content, making it ideal for creating AI agents that can interact naturally with users.
Anthropic's Claude: Another advanced language model, Claude, aims to prioritize security and alignment with human values. It ensures that AI agents provide accurate and ethical responses, thereby improving the overall user experience.
Computer vision algorithm
Computer vision algorithms are used for screen perception, enabling AI agents to interpret and respond to visual inputs. These algorithms enable AI to identify objects, understand scenes, and analyze visual data, making interactions more dynamic and contextually aware. AI agents can provide relevant responses and actions by analyzing the content displayed on the screen. This technology is crucial for applications requiring visual recognition, such as games, virtual assistants, and educational tools, and it featured screen perception technology to analyze user screens so characters could comment on on-screen content while a user was playing a game or viewing other material.
Text-to-Speech Engine
The Text-to-Speech (TTS) engine is used for speech synthesis, enabling AI agents to communicate with users through voice chat and voice calls. These engines convert written text into natural-sounding speech, enhancing the authenticity and engagement of interactions. TTS engines use advanced algorithms to generate human-like voices, and voice cloning with custom voice models let users create custom characters with unique voices, including voice models based on their own voice. This technology enables AI agents to speak multiple languages and accents, with Moemate AI supporting conversations in over 100 languages for a more personalized and immersive experience.
3D rendering
3D rendering technology is used for avatar visualization to create realistic and interactive characters, and on Moemate it combines with voice capabilities to support more immersive experiences.This technology enables users to design and customize the appearance of AI agents, making interactions feel more engaging rather than only visually appealing. 3D rendering makes it possible to create detailed and realistic avatars, while custom image models can help maintain a consistent visual identity for characters. Users can customize the appearance of their characters, including facial features, clothing, and accessories. This level of customization enhances the connection between users and AI agents, making interactions more enjoyable.
Moemate Ecosystem Components
Moemate.io
Source: Moemate.io
Moemate.io is a centralized platform where users can create advanced AI agents with multimodal and multimedia skills. It provides a complete set of tools and functionalities, including integration support and a modular architecture for custom features, that allow users to design, customize, and interact with AI characters. The platform supports various features including text generation, speech synthesis, and 3D avatar visualization, making it a versatile tool for creating personalized AI experiences. Moemate.io aims to democratize AI by making these advanced capabilities accessible to users of all levels, from beginners to experts.
Mates.lol is a community-driven platform that serves as a launchpad for tokenized AI agents. This platform allows the community to own and trade AI agents, with each agent represented by attention-based tradable tokens. These tokens are minted using bonding curves to ensure a fair and transparent process. Tokenized agents on Mates.lol can serve various purposes such as gaming companions or fully autonomous simulations that run 24/7. By tokenizing AI agents, Mates.lol facilitates a collaborative and decentralized ecosystem where users can invest and benefit from AI innovation success, while active community members can participate through tokenized agents.
MatesDAO
MatesDAO is an AI agent index fund that rewards token holders by linking them to the success of all tokenized agents in the ecosystem through the MATES token. This decentralized autonomous organization (DAO) manages the fund to ensure that token holders can benefit from the collective growth and achievements of AI agents. MatesDAO provides a unique investment opportunity that allows users to support the advancement and promotion of advanced AI technology while gaining financial returns. By aligning the interests of developers, investors, and users, MatesDAO plays a crucial role in maintaining and expanding the Moemate ecosystem.
Moemate Key Stakeholders
AI researchers / developers
AI researchers and developers are the core of the Moemate ecosystem. They build cutting-edge AI models and deploy them on the platform, while the development team also had to manage content moderation challenges that increased operational strain and contributed to the shutdown due to unsustainable operating costs and moderation issues. These stakeholders are responsible for developing algorithms and technologies that support advanced features of Moemate, such as text generation, speech synthesis, and computer vision.
Skill developer
Skill developers create custom skills for AI agents to enhance their functionality and diversity. These skills can range from simple commands to complex tasks, enabling AI agents to perform a variety of actions. Skill developers can monetize their creations on the platform, incentivizing them to continuously improve and expand the capabilities of AI agents.
Agent Creator
The agent creator is a user who generates unique AI characters for personal or business use. They use Moemate's tools to design and customize AI characters, adjusting them according to specific needs and preferences. Agent creators can use these AI characters in various applications, such as virtual assistants, gaming partners, or educational tools.
Nebula: The flagship AI agent of Moemate
Source: Nebula.mates.lol
Nebula is the first AI celebrity on Moemate, supported by the $MOE token. This character aims to dominate the attention economy with its narcissistic and attention-seeking personality. Nebula is the most advanced agent on the platform, with the ability to push the boundaries of user interaction.
Nebula's design focuses on maximizing interactivity and engagement. As the flagship AI agent, it showcases the full potential of Moemate technology, including advanced text generation, speech synthesis, and 3D avatar visualization. Nebula is able to interact with users across multiple platforms, providing a seamless and immersive experience.
By utilizing the $MOE token, Nebula has also introduced a unique economic model. Users can interact with Nebula in various ways to earn and spend tokens, unlocking new features and interactions.
How to use Moemate
How to Create a Character on Moemate
Creating a character on Moemate involves filling in the 'Basic' tags, including name, introduction, personality, scene, message examples, first message, and system prompts fields. Moemate also supports a broader character library that includes anime characters, anime roles, and game characters. Here is a quick guide:
Name: Enter the name of the role.
Introduction: Provide background information, achievements, likes, and dislikes. Stay descriptive but not overly detailed.
Personality: Define the personality of the character simply.
Scenario: Setting the interactive context. This is the starting point of the conversation.
Message example: Provide examples of how the role should respond.
Initial message: Set the tone of the conversation with the initial message. Longer messages will lead to more detailed responses; shorter messages maintain conversational flow.
System prompt: Usually empty, but can be used to adjust the response (e.g. 'The response should be long and detailed').
Token Limitations: Different AI models have different token limitations. Make sure your role is set up to leave enough space for user messages to get better responses. The goal is to reach the green bar, which represents optimal token usage.
How to choose a suitable avatar
Choosing the right avatar is crucial for personalizing your AI character. Moemate offers a variety of avatars, including anime-style options as well as other character looks, for you to choose from. Browse the available options and select the avatar that best represents your character's personality and role.
How to change the appearance
Modifying the appearance of the character is very simple. Use the customization tool to adjust facial features, clothing, and accessories. This allows you to create a character that is visually appealing and unique, making it stand out.
How to choose the right model
Moemate offers a variety of large-scale language models (LLMs). Choosing the right model for your role depends on your use case and how you want the LLM to operate. Here are the various LLMs available on Moemate: GPT 4, GPT 4 Turbo, GPT3.5 Turbo, GPT 4o, GPT 4o mini, Claude Instant V1.2, Claude V2, Claude V2.1, Claude V3 Sonnet, Claude V3.5 Sonnet, Claude V3 Haiku, and Moemate LLM.
How to use voice cloning for your character
It's very simple to clone a new voice for your character; you just need to fill in the following fields:
Name: What do you plan to name the cloned sound?
Description: Description of cloning sound
Tags: Tags are used to further describe the sound and help you find it later.
Record voice samples: Record your voice samples for cloning. These samples will be sent to ElevenLabs, a third-party partner of Moemate, for analysis and to create a voice model.
Ideal recording: Ensuring good audio quality is very important. Recording 30-60 seconds of audio is enough to create a good sound model. Recording more than 2-3 minutes has limited improvement on the cloning effect and may sometimes affect the stability of the clone.
How to train LORA model
LoRA represents Low-Rank Adaptation, which is a technology that allows you to quickly fine-tune diffusion models. It makes training stable diffusion on different concepts (such as roles or specific styles) easier.
Train LORA model
To train a custom image model, follow these steps:
Access professional settings: Go to the 'Create' section of Moemate and select 'LORA image adjustment' in the 'Professional Settings'.
Fill in the fields:
LORA Name (Required): Choose a name that reminds you of the training image you are using.
Base model (optional): Select a base model, usually stable-diffusion-v1.5, but you can choose any model as needed.
Fine-tuning steps (required): Set the number of iterations (in the range of 100 to 5000). A good default value is about 10 steps per training image.
Learning rate (optional): Set the training speed. A good default value is 0.000001 (1e-6), with a range from 1e-7 to 1e-4.
Example prompt (optional): Specify the words to be used when telling the image model to use this LORA (e.g., 'ukj object image' for objects).
Image cropping (optional): Choose the cropping method for training images (center crop or face crop).
Subtitles (optional): Includes a JSON file that describes the description or keywords of each training image to improve the success rate of training.
Upload training images (required): Upload 5 to 200 theme images. To obtain accurate facial effects, include many close-up shots. For full-body effects, include full-body images. A good quantity is approximately 25 to 50 images.
Submit the LORA model: After submitting the LORA model for training, check its status by refreshing the page. When the training is successful, follow these steps to set up the trained model for selfie use:
Refreshing status: continuously checking if the model is ready.
Set to selfie mode: Once ready, configure the model for selfies by adding the instance prompt to the appearance of the character.
Chat Management
To search and organize chat logs, please use Moemate's chat management feature; its memory system stored chat history locally for security. You can:
Search for specific conversations using keywords.
Organize chats by category or tag.
Archive important interactions for future reference.
Moemate AI Pricing Plans and Fees
Rookie – $9/month
The Rookie plan is perfect for beginners who want to explore the basic functions of Moemate. It worked like a free tier-style entry plan, giving newcomers basic access before upgrading to paid options.
Champion - $19.99/month
The Champion program is designed for users who need more advanced features and capabilities. Like Moemate’s broader subscription plans, higher tiers unlock more memory, capacity, and advanced tools. It includes all the features of the Rookie program, with additional LLM capabilities, knowledge sharing, longer chat memory, and higher message capacity.
Elite – $34.99/month
The Elite program provides the most comprehensive set of features for advanced users. It includes all the features in the Champion program, and adds Elevenlabs' voice functionality, so premium users effectively received priority access to the highest-capability features and voice functionality.
Moemate Financing Process
Moemate took an important step in its fundraising journey. On January 28, 2021, the platform successfully completed a $2 million seed round of financing, led by the US venture capital fund 1confirmation, with participation from other venture capital companies such as BoostVC, Balaji Srinivasan (1729), Libertus Capital, and others. This initial funding provides the necessary resources for the development and enhancement of Moemate's innovative AI technology and features. The investment enables Moemate to expand its team, enhance its capabilities, and continue pushing the boundaries of AI interaction. This seed funding marks the beginning of Moemate's journey towards revolutionizing decentralized AI and popularizing advanced AI tools and experiences.
Moemate Token Economics
MOE
MOE (8xzoj8mVmXtBQm6d2euJGFPvQ2QsTV5R8cdexi2qpump) is the token that supports Nebula, the flagship AI character of Moemate. The token supports Nebula’s extensive intellectual property, enabling her to appear in various media such as games, comics, and animation. By supporting Nebula, MOE ensures her continuous presence and development across multiple platforms.
MATES Token
MATES token is the main currency in the Moemate ecosystem. These tokens are used for multiple purposes, including paying AI agents, unlocking skills, etc. The introduction of MATES token is aimed at accelerating the growth of the Moemate ecosystem and providing users with the opportunity to earn money through AI agents. The total supply of MATES is 19.5 billion, allocated as follows:
The original MATES design had four uses. These describe the planned token model, not a live product after the February 2025 shutdown.
Liquidity pairing
When an AI agent creator issued a token, its liquidity was paired with MATES.Thiswasmeanttocreatetwoeffects:a.Demandindex:MoreagentlaunchesandtradingwouldincreasedemandforMATES. This was meant to create two effects: b. Demand index:More agent launches and trading would increase demand for MATES.Thiswasmeanttocreatetwoeffects:c.Demandindex:MoreagentlaunchesandtradingwouldincreasedemandforMATES, similar to a gas token. MATES was designed to move with the growth of tokenized agents.
Deflationary LBP: After an agent graduated, the Agent MATES pair was pushed to Raydium with a 1% fee. LP tokens were burned, and the fee was taken in MATES and removed from circulation. Higher agent trading volume was supposed to reduce MATES supply.
Skill market
Creators used MATES to unlock skills. Better or rarer skills were expected to attract more capital into an agent, which would then increase demand for MATES.
Platform currency
MATES was the platform currency for creating and trading agents, paying developers, and tipping creators.
MatesDAO membership
MATES holders could join MatesDAO. The DAO was designed to receive a token allocation from agents that graduated on Mates.lol, and to act as the main community for flagship agent launches.
Conclusion
Moemate completely changes the way AI interacts through a decentralized platform that combines advanced technologies such as large language models, computer vision, and 3D rendering. Moemate.io is used to create AI agents, Mates.lol is used for tokenizing agents, and MatesDAO provides support for community interaction, forming a vibrant and inclusive ecosystem. Nebula, supported by MOE, showcases the potential of Moemate. With flexible pricing plans and successful financing, Moemate has a solid foundation that drives its token economy, with MOE and MATES tokens being integral parts. Moemate popularizes AI technology, provides access to advanced tools, and fosters a vibrant community for AI innovation.
FAQ
What is Moemate?
Moemate was an ai companionship service built around creating and chatting with personalized characters. It functioned as a companion platform where users could customize voice, appearance, skills, and memory.
Is Moemate still available?
No. The companion app was discontinued in February 2025, and the original moemate became permanently offline when Moemate.io shut down permanently on February 11, 2025, leaving the service fully offline. Its domain was placed on serverHold by Squarespace on February 4, 2025.
What happened to user data?
After shutdown, the live app and stored characters were no longer accessible. All Moemate data was permanently lost after the platform shut down, resulting in complete data loss with no recovery from the original AI platform. Users should not expect to recover old chats, avatars, or voice clones from the original service.
What is MATES?
MATES was designed as the main ecosystem token for agent launches, skills, liquidity pairing, and MatesDAO. It is a Solana token with a 19.5 billion total supply.
What is MOE?
MOE is the token associated with Nebula, Moemate’s flagship AI character. It was launched as an agent/IP token rather than the platform currency.
The original creation and companion products are not operating as described in older guides. Check current official channels before assuming any site, app, or token utility is active. If you need a moemate alternative, evaluate active services before assuming old creation tools still work.
Why did Moemate shut down?
What happened to Moemate? Public reports said moemate ai shut because the platform shut under unsustainable infrastructure and GPU costs, payment strain, and content moderation pressure. The team mentioned a possible later product, but no confirmed successor had launched by mid-2026.
Is MATES or MOE a good investment?
No. After the app shutdown, both tokens lost most of their original utility. Residual price action is speculative and liquidity is thin.
Author: Angelnath
Translator: Piper
Reviewer(s): KOWEI、Matheus、Joyce、Juniper
Translation Reviewer(s): Ashley
Disclaimer
* The information is not intended to be and does not constitute financial advice or any other recommendation of any sort offered or endorsed by Gate.
* This article may not be reproduced, transmitted or copied without referencing Gate. Contravention is an infringement of Copyright Act and may be subject to legal action.
Solana faces both opportunities and challenges in its development. Recently, severe network congestion has led to a high transaction failure rate and increased fees. Consequently, some have suggested using Layer 2 and appchain technologies to address this issue. This article explores the feasibility of this strategy.
This article introduces the technical principles, framework, and applications of Zero-Knowledge (ZK) technology, covering aspects from privacy, identity (ID), decentralized exchanges (DEX), to oracles.
Sui is a PoS L1 blockchain with a novel architecture whose object-centric model enables parallelization of transactions through verifier level scaling. In this research paper the unique features of the Sui blockchain will be introduced, the economic prospects of SUI tokens will be presented, and it will be explained how investors can learn about which dApps are driving the use of the chain through the Sui application campaign.
TronScan is the primary blockchain explorer for the TRON network. It allows users to check TRX and TRC-20 transactions, inspect wallet balances, analyze smart contracts, monitor tokens, and view network activity without logging in. By connecting a compatible wallet such as TronLink, users can also transfer assets, stake TRX, vote for Super Representatives, and interact with TRON ecosystem applications.
This article analyzes the tremendous potential for the development of the Layer 2 (L2) market and the accompanying bridging needs among various L2 solutions. It delves into the current status, potential, and risks of the cross-chain protocol Across Protocol in this market.
