ZeroAI LFG application

Hello, JUP community!
This is the ZeroAI team, an innovative initiative focused on predicting cryptocurrency prices using artificial intelligence. Our platform leverages advanced machine learning algorithms to analyze vast datasets, offering insights into future price trends. Exciting times lie ahead with ZeroAI! :laughing: :revolving_hearts: :two_hearts:

What’s ZeroAI?
Envision our platform as CoinMarketCap (CMC), enhanced with AI-driven price predictions. Upon the completion of token integration, users will have access not only to real-time prices and essential information but also to AI-generated forecasts for prices one hour and one day ahead.
Our primary goal is to disrupt the traditional notion that only humans can formulate effective price strategies. Through the implementation of AI-driven predictions on the top CEX and DEX platforms, we seek to democratize access to advanced analytics and empower users with data-driven insights.

How the product collaborates with the Jupiter & Solana ecosystem

2024 is the time when CEX and DEX order books display not just the current price but also predictions for one hour and 24 hours ahead, as forecasted by AI.

Why not offer users more options? In a scenario where AI competes with human KOLs for accuracy in predictions, who will prevail? AI doesn’t necessarily have to win outright, but even capturing 50% of the prediction market—currently dominated by humans—would be a significant milestone.

Jupiter, a decentralized exchange with perpetual functionality, stands at the forefront of innovation in the crypto space. By incorporating AI predictions into its platform, Jupiter can revolutionize the trading experience for users, offering enhanced insights and decision-making capabilities.

Benefits of AI Integration:

  1. Improved Trading Strategies: AI predictions can analyze market trends and patterns, providing users on Jupiter with valuable insights to optimize their trading strategies.

  2. Enhanced Risk Management: By leveraging AI forecasts, Jupiter users can make more informed decisions, mitigating risks and maximizing potential returns.

  3. Real-time Market Analysis: AI-powered predictions offer real-time market analysis, enabling users to stay ahead of market movements and capitalize on opportunities swiftly.

ZeroAI Product
Our Telegram bot and API is equipped to automatically send hourly predictions

Our daily prediction report and weekly report will automatically send by the bot.

What AI algorithm is ZeroAI using?

Core Technologies

  1. Long Short-Term Memory (LSTM) Networks:

These are a special kind of recurrent neural network (RNN) capable of learning long-term dependencies. In our project, LSTM is instrumental in understanding the time-series data of cryptocurrencies, allowing us to capture patterns over extended periods.

  1. Neural Prophet:

This is a cutting-edge forecasting tool that combines components from ARIMA (AutoRegressive Integrated Moving Average) models, Recurrent Neural Networks, and additional regressors. It provides an excellent balance between model simplicity and forecasting power, particularly for seasonal trends.

  1. Random Forests:

An ensemble learning method known for its robustness and versatility. We use it to predict cryptocurrency prices by analyzing a multitude of decision trees and determining the average prediction from various models, thus increasing the accuracy.


The project operates in two main stages:

  1. Initial Prediction Models: Each technology (LSTM, Neural Prophet, and Random Forest) independently predicts daily and hourly cryptocurrency prices. These models are fine-tuned through extensive testing and hyperparameter optimization.

  2. Meta-Modeling Using Machine Learning: We then synthesize these individual predictions, comparing them with historical real-world data. The machine learning meta-model harmonizes these diverse predictions into a single, more accurate forecast.

ZeroAI’s AI algorithm code is completely open source. We encourage more individuals to join us in enhancing our computational models.

Team member
Our team consists of five members, with two core developers possessing a background in AI. While our team remains anonymous for the time being, we have chosen this approach to shift the focus onto our code and algorithm, and that’s why the code is fully open-sourced. By prioritizing the quality and integrity of our technology, we aim to demonstrate our commitment to delivering a robust and innovative product. Our decision to remain anonymous underscores our dedication to transparency through code scrutiny, allowing our work to speak for itself.

Vision of the project

  1. If ZeroAI were to succeed, it would bring a unique offering to the Web3 space by providing AI-driven price predictions for cryptocurrencies, enhancing the accuracy and efficiency of trading decisions for users. This could potentially revolutionize how investors and traders interact with the crypto market, leading to more informed and profitable decision-making processes.

  2. The biggest challenge or obstacle for ZeroAI as a project could be gaining widespread adoption and trust within the crypto community. Establishing credibility, building a strong user base, and continuously improving the accuracy of price predictions will be crucial in overcoming this challenge.

  3. Advice for another team launching a token in Web3 would be to focus on creating real value for users and building a strong community around the project. Transparency, communication, and delivering on promises are key factors in gaining trust and support from the community.

  4. Our future plan is driven by the vision to transform the current landscape, where price strategies are solely reliant on human input. By collaborating with leading centralized exchanges (CEX) and decentralized exchanges (DEX) to integrate AI price predictions, we aim to revolutionize how users approach cryptocurrency trading

Our partner:
CEX: Salavi exchange, CMC top 100
Salavi Exchange has integrated our API, and more CEXs are in discussion.

What’s the Token function?

The tokens will function as a governance mechanism, enabling holders to actively participate in decision-making processes. This includes voting on which tokens should be added for prediction and selecting partners to utilize our API.

Expanding utility for users: In addition to enhancing trading strategies, our AI price predictions will extend beyond traditional trading activities. Users can leverage AI forecasts for various purposes, including:

  1. Perpetual Leverage: Empowering users to optimize leverage positions and manage risk more effectively based on AI-driven market predictions.
  2. Price Betting in Game Pools: Introducing a new dimension to cryptocurrency engagement, users can participate in price betting games utilizing AI forecasts for informed decision-making.

Additionally, all benefits generated by the AI algorithm will be fairly distributed among the entire community of token holders.

As a project focused on product development, we believe in launching our token at the right time rather than rushing it. Jupiter LFG is our chosen partner for collaboration. IT’S THE TIME!!!:muscle::fireworks::fire!!! :


The ZEROAI token launch is on the horizon, and we’re open to discussions with prominent mainnet ecosystems. Contact us: Website: https://0AI.AI


Telegram: Telegram: Contact @ZEROAI_channel

Telegram bot: Telegram: Contact @zero_aiBot

Github: ZEROAI · GitHub


Free API: API Documentation - 0AI.AI


Interesting, but would you open prediction for more tokens and how long would it take? And if I want to apply for my token, any specific conditions??

1 Like

ZeroAI will soon add the top 100 tokens listed on CoinMarketCap for price prediction. Following this milestone, we plan to expand our service to include a wider range of tokens.


This sounds incredibly cool and really excited to see where you can take this! How do you plan on refining the model further? and have you looked at adding different types of data to the set? I think some kind of social media sentiment could be interesting for the model.

1 Like


Let’s say 0AI bot makes 100 predictions, out of 100 how many predictions were accurate?
Can you please share this data if you have it?


The hightlight parts are very close to the real price. Of course, you can follow our bots to get the prediction.


Thank you for your enthusiasm! We are thrilled by the possibilities ahead. To enhance our model further, we intend to refine it through continuous iteration and feedback mechanisms.


Hello there.

Epic introduction. Thanks.

That would be a fire integration for the Jup Perpetuals where you could toggle AI prediction if you wish imho.

How would you try to gain more Users? A Paradigm shift will take a bit. How will you incentivize people to use it?

A possibility could be if a price prediction goes a certain percentage wrong he she could get your Token for the loss (with a toggle if he she wants to )

Only a thought from the top of my head.

I would like to see ya grow




Integrating AI prediction into Jup Perpetuals would indeed be a groundbreaking development, offering users the option to leverage advanced technology for more informed decision-making.

In terms of gaining more users, ZeroAI aims to implement a user-centric approach by prioritizing transparency, accuracy, and user incentives. To incentivize adoption, we plan to introduce a unique feature where users have the option to receive ZeroAI tokens in the event that a price prediction deviates beyond a certain threshold. Of course, more ideas will be considers before token launch with Jupiter LFG


A possibility could be if a price prediction goes a certain percentage wrong he she could get your Token for the loss (with a toggle if he she wants to )

We aim to create a mutually beneficial ecosystem where users feel empowered and rewarded for participating. We did think about compensating users who suffer a loss with AI prediction, that’s part of the roadmap.


I am excited for you guys and our future together

1 Like

thanks sir. We will do our best

Just noticed that your ETH data-set is daily and hourly prices dating back to 2018. And most of your models used are essentially basic machine learning models. On first glance these seem like hardly enough to train any ground-breaking models.

Is there somewhere that you describe your approach in more detail? Is there some novel approach you are taking that’s innovative or different from all the others?

AI improvements generally requires a lot of research, otherwise it’s hard to see how you can beat even the existing deep-learning/LLMs AI models from the tech giants. Could you share at least some background of your team’s AI/ML research capabilities? Thank you.


Yes, our initial models are based on LSTM, Neural Prophet, and Random Forest. We’ve indeed experimented with a variety of algorithms, including ARIMA and XGBoost, among others. After continuous comparative testing and hyperparameter optimization, we chose the outputs from these three models as our seed.

Following this, we trained a meta-model using both the seed results and real data to generate our final predictions. The rationale behind this approach is to mitigate the errors inherent in single models, thus bringing our predictions closer to reality. To our knowledge, we are the pioneers in utilizing this stacked model architecture in our niche domain.

Cryptocurrency prices are influenced by a myriad of factors, including market trends and policy changes. Our current phase focuses on datasets that start with time series and various economic indicators, which have shown promising results within controlled cycles. Moving forward, we plan to incorporate additional non-economic indicators, requiring extensive data filtering, annotation, and correlation.

Drawing a parallel with’s success in the role-playing chat model domain, it’s clear that specialization within the vast field of large language models can lead to significant advancements. ZeroAI is dedicated to leading the field in cryptocurrency prediction.

Regarding your question about our team’s background, I can share that our lead members include a Ph.D. in computer vision from Stanford, along with several engineers with extensive experience in AI. As our project is still in its early stages, we’re keeping the full team details confidential until closer to our official launch.

Thank you for your interest and support.

1 Like

Hey guys love this idea and will be a real hit. Got me think… doesnt black rock have Aladdin that dominates the stock market ? Cant something similar be developed in this space ?

1 Like

This is a fantastic suggestion. The inevitability of AI taking over asset and risk assessment is clear. Our project, at its current stage, represents only a small part of what the future holds, with AI-driven predictions signaling the way forward. In fact, we are in the process of developing and perfecting a number of small extension components, including browser extensions, Telegram bots, APIs, etc., which will provide users with valuable information based on AI reasoning and data integration. Simultaneously, we are working on improving AI algorithms and stacking models, which will be updated on our website and GitHub repository. Thank you once again for your suggestion. We hope to stand shoulder to shoulder with Aladdin in the cryptocurrency domain. :slight_smile:

1 Like

Using AI to create an indicator is pretty interesting. What is your moat?

1 Like

Sounds cool but how are you gonna generate revenue?

1 Like

Would users deposit funds into a zeroai wallet like Drift protocol does? could users deposit into a pooled order to split gains based on % i think that would increase model predictably and reduce losses in that event? And what are the fees like?

New to ZeroAI but so far I really like it, can’t wait to see what you build :raised_hands: