# Introduction Quickly navigate between the best AI models and rate their performance for crypto rewards ## Welcome to Aggregated AI "Aggre" Aggre rolls up the best AI models into one app under a single subscription; offering the optimal workflow for tasks that require multiple models. #### You can download the Aggre app now on the [Apple App Store](https://apps.apple.com/us/app/aggre/id6474451273) and [Google Play Store](https://play.google.com/store/apps/details?id=com.babysamo.babyai\&pcampaignid=web_share)!
## How Aggre Works ### Use Case A: [Swipe Between AI Models Efficiently](/overview/swipe-between-ai-models-efficiently) 1. Quickly navigate between different AI models from a single mobile app experience under one subscription 2. Interface with one AI model at a time and swipe between models based on the task at hand ### Use Case B: [Use Many AI Models at Once](/overview/use-many-ai-models-at-once) 1. Submit a prompt via Aggre's chat interface 2. Aggre routes the prompt through the APIs of top AI models like GPT-4, Claude v2 and Perplexity 3. Aggre answers the prompt with responses from the top AI models 4. Rate which AI model provided the best answer in exchange for a crypto reward (AG token). *This is possible because the data being generated is **really** valuable!* # Supported AI Models Status of Aggre's AI model integrations ### Active Integrations :white\_check\_mark: GPT-3.5 :white\_check\_mark: Perplexity :white\_check\_mark: Claude v2 ### Pending Integrations :yellow\_square: GPT-4o :yellow\_square: Grok :yellow\_square: Llama3 ### Upcoming Integrations :orange\_square: Bard # Goals ## Current Goals 1. Make it easy to navigate between the best AI models on a mobile device 2. Give users access to the best AI models under one app and one subscription 3. Leverage the value of user data contributions to unlock FreeAI for users
## Long Term Goals 1. Make open source AI models as accessible as closed sourced competitors 2. Custody open source AI models on a decentralized AI cloud servers to contribute towards sovereignty 3. Train the [Aggre Local LLM](/upcoming/aggre-local-llm) on in-house data # Team The Neko Fdn is developing Aggregated AI. Neko started on NEAR Protocol in December of 2021 and our community has grown A LOT over the years. We launched NEKO, the first and leading meme coin on NEAR held by over 30,000 unique wallets. Neko has trailblazed NEAR-native content creation and developed a lot of cool stuff along the way too. # FAQ #### **How do you pronounce "Aggre"?** "Aggre" is pronounced as /ˈɛɡrə/ (Egg-ruh). #### What is Aggre? Aggre is a platform that allows users to access multiple top AI models through a single subscription. Users can switch between different AI models based on their specific needs, ensuring they use the best tool for each task. #### Can I use Aggre on my mobile device? Yes! Aggre is available on iOS and Android devices via the Apple App Store and Google Play Store. #### **How does Aggre work?** Aggre routes user prompts through the APIs of various AI models. Users can swipe between models within the app, accessing different AI services without the need for multiple subscriptions. #### **Why should I use Aggre instead of subscribing to individual AI models?** Aggre offers a cost-effective solution by basing subscription prices on the cost of underlying AI credits, avoiding the high premiums charged by individual providers. It also simplifies the user experience by consolidating multiple AI models into one app. #### **How do I earn AG tokens?** You earn AG tokens by submitting prompts and rating the performance of AI models. The more you engage with the platform, the more AG tokens you can earn. #### **How much does an Aggre subscription cost?** Subscription costs vary based on the tier and quantity of AI credits you choose. Our pricing is designed to be competitive and more affordable than subscribing to multiple AI models individually. # Links
ItemDescriptionLink
WebsiteAggre official websiteaggreagted.app
Apple App StoreDownload Aggre on the Apple App StoreAggre iOS
Google Play StoreDownload Aggre on the Google Play StoreAggre Android
TwitterAggre official twitter account@aggregated_app
Telegram
# Swipe Between AI Models Efficiently Aggre optimizes work flows that involve the use of multiple AI models ## Why Use Different AI Models? Leading AI models roll out updates regularly and new specialized models enter the market all the time. This has created a dynamic and fiercely competitive environment. The result? Varied performance between AI models across different subjects and tasks. Aggre lets users pick the right tool for the job *painlessly*. ## Faster AI Workflows Jumping between different app interfaces and logins is really cumbersome *(especially on mobile).* Aggre eliminates much of this friction. Navigating between AI models is as simple as swiping left or right in the Aggre app. Use one model with the others on standby. ## Less Money, More AI Companies like OpenAI charge premiums for subscriptions to turnkey AI models like GPT-4. When subscribing to multiple AI models, it is easy to rack up $60 per month or more in costs. Aggre prices its subscriptions based on the cost of AI credits spent interfacing with the APIs of top AI models. There are no high premiums to use these APIs, so Aggre users bypass the extra costs. Users can decide what AI models they allocate their AI credits to. # Use Many AI Models at Once Instead of relying on a single AI model, Aggre let's you see what all of the top AI models have to say Users submit their prompt and Aggre routes it through the APIs of top AI models. Aggre responds with answers from every AI model, simultaneously.
### Imagine you are trying to solve a problem... #### Would you rather sit in a room with one genius and ask them questions? Or, sit in a room with a dozen geniuses and ask them all questions? *If you chose a dozen geniuses, Aggre is definitely for you!* With responses from multiple AI models, user's gain a broader perspective on the optimal answer. Detecting hallucinations is easier too, with outlier responses standing out from the rest. ## Rate the Best Response, Earn Crypto After receiving responses to a prompt from the top AI models, users can rate the best response in exchange for AG token rewards. This action generates valuable data on the real world performance of competing AI models which can be packaged up and sold. The AG token rewards compensate users for this data and the extra time spent reviewing multiple prompt responses. :point\_right: Learn more about [AG Token](/overview/ag-token)! # Aggre Data An example of data generated by Aggregated AI ## Data Generated by Users See below for an example of the components that make up Aggre "data points".
### User Prompt "Tell me about the state of quantum computing in 100 words or less" ### GPT-3.5 response Quantum computing is an emerging field that leverages quantum mechanics to perform complex calculations at speeds far beyond traditional computers. While still in its early stages, advancements are being made by companies like IBM, Google, and Microsoft. Quantum computers have the potential to revolutionize industries such as cryptography, drug discovery, and artificial intelligence. Challenges remain in scaling up the technology, reducing error rates, and making it more accessible. Despite these hurdles, the future of quantum computing looks promising, with ongoing research and development pushing the boundaries of what is possible in computing. ### Perplexity response Quantum computing leverages quantum mechanical phenomena like superposition and entanglement to perform calculations exponentially faster than classical computers. However, it faces significant challenges, including decoherence and the need for scalable, high-quality qubits. Despite these hurdles, significant progress has been made, with companies like Google and IBM developing and testing quantum computers. These machines have already demonstrated quantum advantage, though practical applications remain limited due to current technological constraints. ### Claude response Quantum computing is a rapidly evolving filed that promises to revolutionize computing power. Unlike classical computers, which use bits as 0s and 1s, quantum computers use quantum bits (qubits) that can exist in superposition, allowing them to perform certain computations exponentially faster. While still in the early stages of development, quantum computers have demonstrated their potential in areas such as cryptography, optimization, and simulation. Major tech companies and research institutions are actively working to overcome the challenges of scalability, error correction, and practical applications. The future of quantum computing holds the promise of breakthroughs in fields from drug discovery to climate modeling. ### User Rating * :black\_small\_square: GPT-3.5 * :black\_small\_square: Perplexity * :white\_check\_mark: Claude # AG Token AG is the native token of Aggre 1 billion AG tokens will be minted on TGE with allocations and vesting as defined on the [Initial Supply & Allocations](/overview/ag-token/initial-supply-and-allocations) page. AG features a dynamic rewards system with a flexible supply that increases in parallel with growing revenues and user activity. Growing revenues also apply deflationary pressures to the AG supply with token buybacks and burns. ## Use Cases * Reward users for rating AI models * Aggre governance * Staking rewards ## Withdrawing In-App Rewards AG tokens earned in the Aggre app are locked until a set quantity is earned. Once the user earns enough tokens, they can withdraw to their blockchain of choice (Solana, NEAR Protocol or Arbitrum). Upon withdrawing from the app, the AG tokens are minted on the blockchain selected by the user. For example, if a user withdraws their tokens to Solana, new AG SPL tokens are minted on-chain and deposited into the users wallet. Similarly, if a user elects to withdraw their tokens to the NEAR Protocol blockchain, new AG NEP-141 tokens will be minted. ## AG Staking AG single-sided staking contracts enable users to lock their AG tokens for fixed durations in exchange for variable yields. More advanced staking and yield generating infrastructure is planned for all blockchains that AG exists on. # Dynamic In-App Rewards ## In-App Token Rewards Aggre's in-app AG token rewards aim to offset user subscription costs while fueling new revenues and in-app engagement. In-app token emissions are a product of in-app revenues from paid subscriptions, sale of data points and free tier advertisements. New AG tokens are minted when users claim their AG tokens on-chain. The corresponding revenues earned go to buy / burn AG token from the open market. ## Dynamic Emissions #### The quantity of in-app rewards per rated prompt are based on these key factors: * AG token price * Cost of user subscription * Maximum quantity of monthly prompts $$ \text{Target Value of AG Tokens per Prompt} \geq \frac{\text{Cost of User Subscription}}{\text{Maximum Quantity of Monthly Prompts}} $$ ## How it works ### **Subscription Matching** When a user subscribes to Aggre they pay a monthly fee. Aggre's rewards system aims to reward AG tokens in an amount worth at least what is paid for in the subscription, or even more. NOTE: Lot's of volatility is to be expected in the early days of AG token. For that reason, users must anticipate fluctuations in the value of in-app rewards. In-app rewards will require a ton of fine-tuning. ### **Earning Tokens** Users earn AG tokens by submitting prompts and rating the best responses from different AI models. More prompts submitted -> More responses rated -> More AG tokens earned ### **Balancing Rewards and Engagement** The total quantity of AG token rewards needed to offset the user's subscription costs are divided by the maximum number of prompts the user can submit within their subscription tier. ### **In-App Reward Ceilings** Every subscription tier comes with a maximum quantity of prompts per month. This maximum is based on the hard costs to use multiple AI model APIs. To earn maximum AG token rewards, users will aim to reach the maximum quantity of prompts allotted by their paid subscription. If desired, users can recharge their AI credits on Aggre if they max out for the month. Remember, the goal here is free AI. Anyone buying an Aggre subscription just to rip through AI prompts in an attempt to amass tons of AG rewards will be disappointed. # Initial Supply & Allocations ## Initial Supply 1 billion AG tokens shall be minted on TGE with allocations as defined below. Future AG tokens will be minted based on in-app emissions backed by robust corresponding revenues. ## Token Allocations
CatagoryInitial Supply (%)QTY Tokens
Team10%100,000,000
Liquidity and Emissions30%300,000,000
Public Sales10%100,000,000
Private Investment20%200,000,000
Treasury10%100,000,000
Airdrop & CommunityTBATBA
AIBS Pre-Token AllocationTBATBA
## Vesting Schedule Team: 10% (100M AG) with a 4-year vesting schedule Investors: 20% (200M AG) with a 2-year vesting schedule Treasury: 10% (100M AG) with a 3-year vesting schedule Liquidity and Emissions: 30% (300M AG) with a 4-year vesting schedule # FreeAI Model Goal: Offset the cost of a user's subscription by providing AG token rewards that match, or even exceed the subscription costs ## Overview Aggre's FreeAI aims to offset the cost of user subscriptions with AG token rewards, making access to the best AI models effectively free. This model balances user incentives with platform revenues and expenses, ensuring financial viability. ## How it Works

New AG tokens are minted in parallel with growing Aggre revenues

### Inputs #### 1. Paid Subscriptions: * Users subscribe to Aggre, paying a monthly fee to access AI credits from the top AI models. * Different subscription tiers offer varying quantities of AI credits, determining the number of prompts users can submit. #### 2. User Data Contributions: * Users actively submit prompts and rate the performance of AI models. * These activities generate valuable data that can be packaged and sold as data points. #### 3. Advertisements: * In-app advertisements displayed to free-tier users. ### Outputs #### 1. Operational Costs: * Routing user prompts through top AI models via API calls. * Infrastructure and other operational expenses #### 2. AG Token Rewards: * Users earn AG token by rating AI model performance, with dynamic emissions based on current token price and the user's subscription tier. ## Achieving Economic Balance #### Subscription Revenue vs. AI Credit Costs: * Subscription fees are set higher than the cost of the corresponding AI credits. This ensures that each subscription generates a positive cash flow for the platform. * For example, if a subscription costs $20 per month, the cost of the AI credits provided should be slightly less, ensuring a margin that contributes to overall revenue. #### Balancing AG Rewards with Sustainability: * The platform limits the number of prompts a user can submit per month based on their subscription tier, ensuring that AG token emissions are controlled. Once users reach their maximum prompt limit, they can reload their credits, generating additional revenue for the platform. * It is OK for Aggre to run at a -PNL when considering the value of AG token rewards distributed while in a growth oriented trajectory. For example, OpenAI runs at a -$14/month loss per user. Intelligent tokenomics and real yield can offset AG token rewards. # Supported Blockchains A painless web3 experience? Can it be done? Probably not yet... but we are working to get as close to painless as possible. To help make the crypto experience as comfortable as possible, AG token and associated smart contracts / programs will live on the top blockchain ecosystems including Solana, NEAR Protocol and Arbitrum. Users can withdraw their AG token rewards onto top blockchains such as Solana, NEAR Protocol and Arbitrum. AG token shall be supported by sufficient liquidity provisions and on-chain staking applications on all chains. Due to the extensive resources needed to achieve this, AG token may be rolled out strategically onto each blockchain. # Spin the Wheel Instead of a fixed (lame) quantity of rewards, Aggre users get the chance to win big every time they play the Spin the Wheel game Everyone loves a good Spin the Wheel game! With Aggre, every time you submit a prompt it goes towards a spin on Aggre's Spin the Wheel game. Users are rewarded with a free spin for every 10 prompts they submit. Users can claim a daily spin on the Aggre app as well.
The Aggre [Referral Program](/overview/referral-program) also leverages our Spin the Wheel game. Every new sign-up with your referral code is a fresh spin on the wheel. # Airdrop & Community Rewards A large percentage of the initial AG token supply is allocated to a token airdrop centered around the Aggre app users and community. This airdrop will help ensure governance control is distributed amongst the most active users and contributors. While also rewarding early community members vital in scaling Aggre broadly. The Aggre airdrop is intended to reward early community members and users. ## The First 500 A major allocation of the AG airdrop shall be awarded to the first 500 paid Aggre subscribers. App MVPs suck; full of bugs and limited features. Initial users are a driving force in the proceeding updates which turn a minimal viable product into a decent app experience. First-movers deserve extensive compensation for the role they play in growth. ## Nekosystem Rewards Over the past 3+ years, Neko has built a robust community within the NEAR Protocol ecosystem. Over 30,000 individual wallets hold NEKO token. Thousands of other wallets hold Neko NFT assets from the Good Fortune Felines collection to the Genesis Plushies. Aggre is the most advanced development initiative that Neko has perused. Aggre has the opportunity to create a a lot of value and the Neko community will be a key beneficiary. To ensure this, a great deal of AG token airdrop rewards are reserved for the Neko community. On-chain Nekosystem assets (NEKO token and Neko NFT collections) will function as multipliers when determining how AG airdrop rewards are distributed. # AIBS Pre-Launch Token Aggre was initially branded as "Baby AI" with a puppy mascot. That's why you may find artifacts of cute doggo logos throughout the Aggregated AI website and app. To facilitate in-app rewards during the initial market testing phase, AIBS token was launched on the Ethereum blockchain. AIBS token was later bridged to the Arbitrum L2 for its lower transaction fees. Preliminary AIBS staking contracts are deployed to Arbitrum as well. AIBS CA: An AIBS <> wETH liquidity pool is live on Uniswap (Arbitrum). This token will act as a pre-allocation to AG token that will be exchanged following TGE. Users can trade via the AIBS <> wETH liquidity pool on Uniswap (Arbitrum) until the AG token is live. # Referral Program COMING SOON Aggre is launching an interactive referral program with instant reward opportunities. Every new referral earns a spin from Aggre's spin the wheel game that features crypto rewards with every spin.
#### More referrals = more spins = more crypto instantly. Referral sign ups are monitored progressively with a potential airdrop multipliers based on how early the user has signed up. # Generators (Sunsetted) Aggre AI generators are sunsetted for now Version 1 of Aggregated AI consisted of 13 AI-powered generators helping users create emails, tweets, tinder bios and more. This marked our first voyage into mobile app and AI development. Aggre has evolved and we have transitioned away from AI generator tools. Aggre generators are available on the app at this time but are not the subject of development. # Aggre Local LLM ## Meet Neko :wave: The Aggre local LLM
#### Neko is our Dell PowerEdge R730xd server sporting dual NVIDIA RTX A4000 GPUs, 44 CPU cores, lots of DDR4 ECC RAM and NVME storage. We finished the initial build in July 2024 and will post more updates on our progress as the year progresses. # Open Source AI Models on Decentralized Cloud Making open source AI models as accessible as their closed source competitors ## Simplifying Access to Open-Source AI

Front page of Hugging Face as of 8/6/2024

Over 800,000 open source AI models are available on Hugging Face today. However, running one of these models locally requires expensive server hardware, a lot of time and some knowledge. This is just not feasible for most people. Running multiple open source AI models means scaling up all associated inputs. It's a lot, even for the pros. Aggre is making open source AI models as easy to use as their popular closed source counterparts. Users hop on the Aggre app and interface with open sourced AI models the same way they interface with ChatGPT: \*submit a prompt\*. Aggre does the hard parts in the background. We train and manage the models. The only thing left to do is use them. ## Using Open Source AI on Aggre Aggre users can purchase a premium subscription to Aggre for access to the top open source AI models. Costs to operate on the AI cloud are funded from premium subscriptions. This specialized subscription is designed for industry professionals, researchers and enthusiasts that want painless anytime access to their favorite open source models. Benchmarking models from a mobile device sounds pretty nice! ## Running on Decentralized AI Cloud Since open source AI models may save the world someday, we feel responsible in contributing towards their sovereignty. For this reason (among many others) Aggre has elected to use decentralized AI cloud to power its supported open source AI models. Aggre is conducting a feasibility studying on potential decentralized compute providers/marketplaces centered around cost, resources and funding programs. ## Specialized and Uncensored Open Source AI Models There is high demand for uncensored general intelligence and specialized niche models. Running on decentralized AI servers, Aggre can be a first mover in making these models easily accessible. # Text-to-Image Generators Aggregating text-based AI models is just the first step for Aggre.