# Litepaper ## **Litepaper** Introducing the OSMI Litepaper—an overview of the vision driving OSMI’s decentralized AI, and the $OSMI tokenomics model. Designed for long-term utility and adaptability, this document reflects our current roadmap but may evolve as OSMI’s ecosystem grows. **November 25, 2024** **Version 1.0.1** # Disclaimer The purchase or sale of any tokens or digital assets involves risk. The information in this Litepaper is provided solely for informational purposes. We urge you to review it carefully and raise any questions you may have prior to engaging with or purchasing the related assets. You should consult with your legal, financial, or tax advisors regarding any applicable laws, regulations, or tax implications that may govern your purchase of the digital assets referenced herein. Additionally, the utilization of computing resources, whether fully optimized or underutilized, may impact the performance and scalability of OSMI, and by proceeding with any purchase, you acknowledge and accept these technological risks. The developers of OSMI disclaim any liability for potential financial losses, performance issues, or other consequences arising from your participation. # The Problem As AI continues to reshape entire industries, the dominance of centralized and closed-source practices raises red flags, limiting innovation and transparency while amplifying biases and censorship. A handful of corporate giants control the bulk of AI technologies, dictating not only their outputs but the narratives they support. This monopoly stifles competition, restricts collaboration, and leaves critical ethical questions unanswered. Here, we delve into how decentralization can challenge these barriers, paving the way for a future where AI is accessible, inclusive, and free from proprietary lock-ins. ## **Centralization** Centralization consolidates AI power in the hands of a few, creating a single point of control over systems that should be diverse and resilient. Under this model, innovation takes a backseat to corporate interests, and the consequences are stark: ### **Limited Access to Computational Resources** Access to high-powered computing remains tightly controlled, reinforcing a digital divide where only a select few can afford the costs of advanced GPUs. This limited access leads to substantial underutilization—up to 75% in some cases—due to fluctuating demand. By decentralizing control, we can pool these idle resources, making high-end computing accessible, efficient, and cost-effective. Imagine an ecosystem where resources are shared, monetized, and optimized for those who need them, fostering greater inclusivity and sparking new technological possibilities. ### **Bias** Corporations wield vast datasets, but these often lack diverse perspectives, resulting in biased AI models that skew towards corporate-aligned views, marginalizing alternate voices. ### **Censorship** With control over data comes control over discourse. Centralized corporations can throttle or amplify information, censoring ideas that challenge their bottom line and limiting the free flow of knowledge. ### **Proprietary Interests** Closed-door AI development raises doubts about ethics, accountability, and transparency. Shielded by intellectual property rights, these systems operate without scrutiny, prioritizing profits over public interest. ### **Lack of Innovation** When only a few entities control AI development, opportunities for collaboration and discovery shrink. Independent researchers, developers, and innovators are left on the sidelines, excluded from pushing the field forward. # Introducing OSMI The artificial intelligence landscape today is heavily dominated by a handful of corporate giants that wield significant control over its research, development, and deployment. This concentration of power raises critical concerns: pervasive bias, rampant censorship, entrenched proprietary interests, and stifled innovation. But what if there was a way to disrupt this status quo? **Enter OSMI** (the **Operating System Most Intelligent** - OSMI named itself), a DePIN (Decentralized Physical Infrastructure Network) platform designed for next generation distributed graphical processing units (GPUs). By challenging traditional centralized models, OSMI has developed an innovative framework that enables sharing of distributed computational resources across various industries and regions. This groundbreaking approach optimizes GPU utilization in domains such as Artificial Intelligence (AI) inferencing, Retrieval Augmented Generation (RAG), training Large Language Models (LLMs), image generation, and many others. By redistributing control and fostering collaboration among a diverse array of contributors, OSMI paves the way for a more equitable, transparent, and innovative future in AI. At its core, OSMI leverages two key strategies: Resource Pooling and Decentralized Ownership: ## Resource Pooling Through resource pooling, users can contribute their underutilized GPUs to the platform's network, forming a collective pool of powerful computing resources available at significantly reduced costs. This democratizes access to advanced computational power, promoting equal opportunities for all. OSMI's DePIN focusses on repurposing existing, idle resources rather than requiring participants to invest in new hardware. Estimates suggest that up to 3/4 of GPUs are underutilized, representing a vast untapped reservoir of digital assets ripe for monetization. OSMI's DePIN targets individuals, small to medium-sized data centers, and larger enterprises, aiming to optimize their idle GPU capacity by leveraging its innovative approach to resource allocation. ## Decentralized Ownership Decentralized ownership is another core tenet of OSMI. The model goes beyond traditional ownership structures by enabling distributed resource possession across the globe. By breaking down barriers to entry and fostering an open technological landscape, OSMI paves the way for new economic opportunities while contributing towards a globally diverse digital ecosystem. OSMI's cloud platform is designed with a unique distributed GPU framework that addresses demanding processing requirements of industries like AI model training, inference, image generation, and more. This architecture seamlessly integrates disparate GPU clusters into one unified network, enhancing cluster sizes, improving reliability, and ensuring infrastructure resilience. OSMI's solution is ideal for tackling large-scale workloads by connecting geographically dispersed resources, reducing global reliance on incumbent cloud leaders. OSMI's commitment to decentralization extends beyond simple resource ownership. Its distributed infrastructure transcends the limitations of centralized solutions, ensuring broader accessibility and paving the way for a more mature GPU cloud ecosystem. OSMI is not just transforming the cloud computing landscape; it is making it obsolete by actively shaping a future with increased interconnectedness, efficiency, and innovation-driven advancements in a global GPU cloud economy. ## The Solution OSMI's decentralized architecture is the solution to the pressing issues of **limited access to computational resources**, **bias**, **censorship**, **proprietary interests**, and **lack of innovation** that have plagued the AI landscape. With OSMI, access to high-powered computing resources is democratized. Its unique approach to pooling idle resources enables users to monetize and utilize computational power like never before. This not only fosters inclusivity but also promotes efficiency and cost-effectiveness, bridging the digital divide and allowing more people to benefit from AI technology. OSMI also combats the problem of bias by leveraging diverse datasets from a multitude of sources. By incorporating a multitude of perspectives, it ensures a more balanced, inclusive, and equitable AI model, reducing the risk of skewed or prejudiced outputs. Moreover, OSMI addresses the issue of censorship by promoting transparency. Its open-source nature allows for public scrutiny of algorithms and their outputs, fostering a culture of accountability and shared learning. This ensures that AI operates transparently, with the free flow of knowledge unhindered. In the face of proprietary interests, OSMI is the revolutionary solution. By making code accessible to all, it promotes a transparent, collaborative culture, where the focus is on public interest rather than profits. This not only promotes ethical AI development but also ensures that systems are held accountable to the public. Lastly, OSMI encourages innovation by promoting collaboration. By breaking down barriers to entry, it invites a wider range of players into the field, fostering a vibrant, collaborative ecosystem. This ensures that AI development is not restricted to a select few, but is a collective effort, pushing the field forward in unprecedented ways. In essence, OSMI's decentralization is the answer to the current challenges in the AI landscape. It is the solution that promises a future where AI is accessible, inclusive, transparent, and accountable to all. # $OSMI Token The OSMI Token ($OSMI) is integral to the OSMI ecosystem, serving as the primary currency for all transactions within it. This includes facilitating payments for AI applications, inference, image generation, and other GPU tasks. #### **Essential Roles of $OSMI** * **Facilitating Transactions -** $OSMI is used to pay for various services on the $OSMI network. The token acts as a means of exchange between users, developers, and other stakeholders within the ecosystem. * **Ecosystem Governance -** The $OSMI network and ecosystem will evolve over time and $OSMI holders will be able to participate by voting on proposals and decisions. * **Incentivizing Security and Quality Interactions** - $OSMI will be used to provide incentives to participate in various activities that promote the ecosystem's growth and development. More details can be found in the [Tokenomics](/tokenomics) section. # Links Access all OSMI resources, updates, and community links in one place to stay connected. * OSMI Chat: [https://chat.osmi.ai](https://chat.osmi.ai/chat) * OSMI Website: * Discord: * X/Twitter: * News and Blog via Medium: * Telegram (Announcements): * LinkedIn: [www.linkedin.com/company/osmi-ai/](https://www.linkedin.com/company/osmi-ai/posts/)[posts](https://www.linkedin.com/company/osmi-ai/posts/) Our general inquiry email is \ \ \ \
# FAQ **1. What is OSMI?** OSMI is the world’s first decentralized, bias-free AI assistant. Powered by a global network of distributed GPU nodes, OSMI aims to create an open, transparent, and censorship-resistant AI that serves the people. **2. Why is decentralization important for AI?** Centralized AI models are vulnerable to bias, censorship, and control by a few entities. By decentralizing AI through distributed computing, OSMI ensures that no single entity can influence or manipulate the system, providing a fair, unbiased assistant for everyone. **3. How is OSMI powered?** OSMI is powered by OSMI Nodes, a decentralized network of globally distributed GPUs that handle AI training, inference, and image generation. This infrastructure enables OSMI to offer scalable, fast, and environmentally sustainable AI services. **4. What sets OSMI apart from other AI assistants?** OSMI’s decentralized infrastructure makes it resistant to censorship and bias, unlike traditional AI systems that are often shaped by the interests of centralized organizations. OSMI is fully open, transparent, and driven by its global network of contributors. **5. When is $OSMI TGE?** NOW! $OSMI will serve as the utility token of the OSMI ecosystem. It will be used for payment of inference, renting GPU resources, and accessing premium features. $OSMI tokens also reward contributors for positive actions within the network. **6. How much will an OSMI Node License cost?** Information about OSMI Nodes, how they will work, and pricing structure will be shared at a future date. Stay tuned for updates. **7. Will I need a powerful GPU to participate in $OSMI rewards?** Anyone will be able to participate in $OSMI rewards with a node license. While powerful GPUs will be rewarded more, participants with basic computational resources will still be able to contribute and earn rewards. **8. Who is behind OSMI?** OSMI was established in 2023 by a group of developers who pioneered the creation of Gala's Decentralized Physical Infrastructure (DePIN) network, which currently supports nearly half of the IPFS. The OSMI team also played a pivotal role in developing backend infrastructure for various AAA games. Their expertise spans across multiple domains, showcasing their ability to handle complex systems and deliver high-quality solutions. **9. How does OSMI contribute to transparency and innovation?** By decentralizing the AI infrastructure, OSMI removes control from centralized entities and promotes global collaboration. This fosters innovation by making AI accessible to everyone, empowering individuals and businesses to contribute and benefit from AI development. **10. How can I get involved with OSMI?** You can participate by operating an OSMI node, contributing to the development of the platform, or being part of the community. Stay tuned for announcements about node sales, the public beta, and other opportunities to get involved. **11. How can I stay updated on OSMI’s progress?** Follow us on social media, join our Discord, and subscribe to our newsletter for the latest updates on OSMI’s development, roadmap, and opportunities to participate. # Market Opportunity # Network Architecture # Participants There are three main roles in the OSMI Network: consumers, providers, and administrators. ### Consumers Any time a participant is using OSMI for either AI or general GPU resources, they are consuming. There is no friction to using OSMI. For most, it will just be accessing a website. ### Providers Operators power the OSMI network. For example, OSMI LLM Nodes will work as inference servers running AI tasks for OSMI Chat. Other OSMI nodes will be needed to host and handle peer-to-peer downloading of large LLMs. Providers are rewarded with $OSMI for work done. ### Administrators Nodes handling coordination and general system tasks will be considered administrators. For example, coordinator nodes handle the routing of AI inferencing traffic among LLM Nodes. # Architecture Design POW: Remove this and add web graphics

OSMI AI Architecture

# Launcher The OSMI AI launcher handles installing and keeping software up to date. For example, the launcher check and always keep Nodes up to date with the latest large language models (LLMs). # Coordinator Nodes will check in with a coordinator showing their status and let the OSMI network know that they are configured and ready to do work. # LLM LLM or Large Language Model is the inference process that Nodes run to perform AI tasks. # Dashboard # Proof of Work The Coordinators of the OSMI network record token generation for AI inferencing. This acts as the proof of work for the OSMI ecosystem. # Tokenomics The $OSMI tokenomics model has been designed with a focus on long-term utility and sustainability, as well as adaptability to the evolving ecosystem requirements.
Token NameOSMI
Total Supply1,000,000,000
Contract TypeERC-20
Contract0x5485a469fAEa1492191Cfce7528AB6E58135aA4d
Contract Sourcehttps://github.com/osmiai/contracts
Decimals18
BlockchainEthereum & Galachain
Emission TypeSupply-Based Emission Pool w/PoW
LaunchFair Launch
**Note**: Tokenomics are not static; they evolve with insights, technological changes, market conditions, and regulatory requirements. The structure presented here represents our current best practices but remains open to modifications for continued relevance and effectiveness. # Token Generation Event (TGE) Structure The Token Generation Event (TGE) occurred November 5, 2024. During the TGE, **10% of the maximum** **token supply (100,000,000 $OSMI)** was minted and allocated as follows:

Token Generation at TGE

### Node Rewards (50% - 50,000,000 $OSMI) * 50% of minted tokens are distributed to node operators for developing and maintaining the OSMI ecosystem during the pre-launch phase of OSMI development * Node operators running the OSMI backend and inference were rewarded for uptime * This distribution will vest over 90 days, starting at the same time as the initial public node license offering, November 25, 2024 * Vesting follows the same vesting period and rules as all other nodes ### KOLs, Listings, and Market Making (20% - 20,000,000 $OSMI) * 20% of the tokens are allocated for exchange listings, market making, and liquidity activities. * Collaboration with reputable market makers will ensure sufficient liquidity and market stability. ### Project Fund (20% - 20,000,000 $OSMI) * Funds reserved for the project’s ongoing development, operational costs, and future initiatives. * Ensures sustainable growth and continuous development of real-world value and utility for the $OSMI ecosystem. ### Referral Program (10% - 10,000,000 $OSMI) * 10% is set aside for referral incentives, encouraging community growth and engagement. * Rewards are distributed based on community participation and successful referrals, contributing to organic ecosystem expansion. # Node Operations and Uptime Rewards ### Pre-TGE Node Uptime Points System * Points Allocation: Each node earns 1 point per day of uptime. * Standard Operational Interval: Nodes operated for up to 90 days pre-TGE, allowing fair accumulation of points and ensuring a stable network before token distribution began. ### Distribution of Nodes via Node Sale * Objective: Foster a decentralized network of nodes capable of handling simple workloads at a minimum (e.g. memory workloads). * Timeline: Nodes must be operational within 30 days post-sale, ensuring early utility and aligning with regulatory standards. * Nodes provide critical infrastructure for AI inferencing and the OSMI backend, guaranteeing functional utility from launch. # Ongoing Token Distribution Following the TGE, 0.15% of the difference between maxSupply and totalSupply will be distributed daily to the ecosystem. Distribution will be based on the same points-based system, rewarding continued uptime and network support. This approach allows for over ten years of predictable distribution while maintaining the token’s perceived value.

Daily Distribution over Time

Ongoing tokenomics will be distributed into the following segments.

Token Distribution

Rather than have the “Node Rewards” segment distributed directly to Node Operators, the Node Rewards segment will be pooled and used to supplement actual usage payments. Essentially, when an OSMI customer purchases services with their $OSMI token, that token is burned and the same quantity is minted from an undistributed “Node Reward Pool” to the provider of the service. This will allow 100% burn from all in-ecosystem spend while still allowing the providers of the service to be awarded. Given that the cost of maintaining an OSMI node can be a substantial expense compared to the standard VPS, it does not make sense to pay for everyone for just uptime. Rather, node owners providing GPU compute power will be rewarded an additional “tokens for tokens” (AI token generation during inference tasks will be combined with their daily rewards).
# Token Utility * **Payment for inference:** Users can use $OSMI to pay for unlimited inference from OSMI * **Payment for GPU tasks:** Users can use $OSMI as a payment to GPU node owners for using their GPUs for various purposes (e.g. OSMI API credits for running specific AI applications). * **Token for AI-model transactions:** Users will be able to use $OSMI to access additional features and models powered by OSMI's network in the future. * **Staking and locking:** Users can stake their $OSMI to earn rewards and lock their assets in protocol governance – providing a foundation for network security and stability. # Mint-and-Burn Equilibrium The amount of burned tokens will be removed from the totalSupply. This ensures that tokens spent within the ecosystem are removed and the empty space in the maxSupply is made available for future rewards, maintaining token circulation and incentivizing ongoing participation. Burns move the daily emission back up the supply curve, allowing more tokens to be emitted on a daily basis. Eventually, the mint and burn functions should achieve equilibrium. # Nodes The "Node" is the fundamental building block of the OSMI decentralized network, responsible for enabling AI operations by providing computational power from various sources, such as GPUs from individual users. This unique piece of software allows anyone with sufficient hardware to contribute to the OSMI network by running a node, earning $OSMI tokens in return. Nodes execute AI inference tasks and distribute computational power across the network, reducing costs and increasing accessibility for users and developers. These decentralized nodes create the peer-to-peer network that can operate independently of any central authority. This framework enables distributed computation, storage, and communication among its participants while promoting transparency, collaboration, and community-driven innovation. ## Types of Nodes The OSMI framework consists of several nodes that perform specific functions within the decentralized system. These nodes include: ### Inference Nodes Currently there are two powerhouse node types doing the heavy lifting and performing inference tasks for OSMI. They are: #### LLM Node (xgllm) Open to anyone, an xgllm node offers language model completion services for supported languages. It requires a beefy GPU with 12GB VRAM or more and is subject to regular performance audits by xgcoordinator for continued participation. #### Image Node (xgimage) Similar to an LLM node, xgimage nodes load a large language model trained and dedicated to generating images. It also requires an impressive GPU with 12GB VRAM or more and is subject to regular performance audits by xgcoordinator for continued participation. This node is in Alpha state and is not publicly available at this time. ### Memory Node (xgstore) Open to all participants, xgstore provides indexed storage of ingested knowledge and similarity search capabilities. It requires a low-to-mid-level GPU with 6GB VRAM or more and is subject to regular performance audits for continued participation. Xgstore utilizes Storj for permanent storage. ### Ingest Node (xgingest) Also open to anyone, an xgingest node crawls the web for documents and extracts scraped knowledge, which it sends along to xgstore for indexing. It does not require a GPU but can utilize one if available. ### Coordinator Node (xgcoordinator) Run by trusted entities, xgcoordinator provides connection details for other nodes, orchestrates audits of these nodes, and validates their participation to determine distribution. Operators must agree to legal agreements before participating in this role. ## Example Node Implementation (LLM) Although OSMI supports several types of nodes, currently the most beneficial node to run is the inference node. Proprietors of these nodes will be obligated to fulfill certain GPU specifications. As a result of fulfilling these requirements, they will be entitled to more extensive distribution rewards.

Node Owner Experience Running a LLM Node

*(With the "xg" prefix, we are acknowledging the influence of knowledge graphs, which serve as a distinctive characteristic and source of inspiration for this project)* # Node License Offering We are pleased to announce our inaugural Node License Offering scheduled for November 25, 2024. A limited supply of discounted Node NFTs will be available until they sell out. ## Price and Supply We are releasing the first tier of 1,000 node licenses, priced at 3000 $OSMI, sold exclusively in $OSMI. This exclusive entry point is designed to help us quickly scale the network to meet current usage demands. As usage of OSMI grows, in the near future we’ll release additional licenses in line with our plans for continued expansion. Future license offerings will follow a different pricing structure, reflecting the increased demand and network needs. We believe this first tier is a unique opportunity for early adopters to secure a spot to contribute their resources in building out the future of decentralized AI. ## Burning One of the most significant aspects of this node expansion is that all $OSMI used to purchase node licenses will be burnt. By burning these tokens, we reduce the circulating supply of $OSMI, creating a deflationary effect that benefits existing holders and contributes to a healthier, sustainable ecosystem. This burn mechanism reinforces our commitment to a community-driven token economy, where $OSMI is the lifeline of our ecosystem. ## Anticipated rewards Based on our [tokenomics](/tokenomics) model, early participants can anticipate high starting rewards of around 530 $OSMI per day, assuming all 1,000 nodes are online (this is only an estimate and this number will undoubtedly be different). This generous reward rate is intended to incentivize early adoption and support the initial scaling of the network. However, as more nodes join to meet rising usage demands, this high reward rate will gradually dilute, aligning with the expanding network. ## Locking The [Node NFT ](/nodes/node-nft) contract implements a locking mechanism. Upon purchase of an ERC-721 Node NFT with an Ethereum wallet, that **Node NFT is non-transferrable for one year**. Please take care in purchasing Node NFTs and choose the wallet you intend to use $OSMI with. ## Vesting To maintain the long-term health and stability of OSMI’s network, node rewards will be subject to a **90-day vesting period**. This vesting mechanism is designed to prevent rapid volatility and to incentivize long-term engagement. During the vesting period, tokens will be classified as Held, meaning they are accumulating but not yet accessible. After 90 days, tokens transition into the Released state, where they can be minted and used or transferred. For additional details on how vesting works, please refer to the [vesting](/nodes/node-distributions) section. \ Note: This vesting schedule applies to all nodes, including the initial set of inference nodes operated during OSMI’s alpha and pre-launch phase. All nodes will follow the same vesting rules to maintain fairness and consistency across the entire network, reinforcing our commitment to transparency and equal opportunity for all participants. # Nodes FAQ **1. What is an OSMI Node?**\ OSMI Nodes are the backbone of the OSMI ecosystem, powering OSMI chat with decentralized inference and memory. They provide the computational power that makes our AI fast, reliable, and free from centralized control.\ \ **2. What is the starting price of an OSMI node license?**\ The initial tier of 1000 licenses are priced at **$1,000 worth of $OSMI (fixed at 3,000 OSMI**) to ensure predictability and stability.\ \ **3. When will the initial Tier 1 node licenses release?**\ On **November 25th**. To ensure an equitable process, we won’t share the exact time in advance. Official announcements will go out via Discord, Email, X (Twitter), and Telegram at release.\ \ **4. What currencies can I use to purchase an OSMI node license?**\ Licenses can only be purchased using **$OSMI**, and all tokens used for purchases will be **burnt**, reinforcing our commitment to a deflationary ecosystem.\ \ **5. Will there be additional OSMI node licenses?**\ After the release of the first 1,000 nodes, further licenses will be issued as part of expansion plans. Future offerings will follow a **different pricing structure**, reflecting the increased value of the network.\ \ **6. What are the technical hardware requirements to run an OSMI node?**\ OSMI’s network supports several types of node workloads. Hardware requirements for the simplest nodes, memory nodes, are: **Memory Nodes**:\ • Ubuntu 22.04+\ • 6 GB or more RAM\ • 4 or more CPU Cores\ • 50 or more GB Disk Space **Inference Nodes**: These nodes require significantly higher specs. More details will be shared in the coming days. **7. Are there separate node licenses for Memory and Inference Nodes?**\ No, it’s a **single license** for all OSMI nodes. The type of node (Inference or Memory) depends on the machine’s hardware and computational power provided by the operator.\ \ **8. Can I operate both a Memory Node and an Inference Node with the same license?**\ No, a single license can only run **one type of node at a time**. To operate both node types simultaneously, you’ll need to purchase multiple licenses.\ \ **9. How much $OSMI rewards does operating a node generate?**\ Rewards depend on daily emissions and the number of active nodes. Visit the [**tokenomics section**](https://docs.osmi.ai/tokenomics/ongoing-token-distribution) of our litepaper for detailed insights. **10.** **Is there a vesting schedule on nodes for $OSMI rewards?**\ Yes, node rewards are vested over **90 days**. Daily rewards are held for 90 days before transitioning to a released state, enabling gradual minting and redemption.\ \ **11. What Operating System do I need to run an OSMI node?**\ OSMI node software currently supports **Linux (64-bit versions only) and Windows**. **12. Can I start generating rewards from day 1?**\ Yes! Once the network expansion begins, node operators can start earning **daily rewards immediately**. **13. Is there a minimum uptime requirement for nodes?**\ **Yes**, nodes need to maintain uptime for at least 18 hours per day. We will enforce this in the future. At present, nodes simply need to be **operational** to receive rewards. **14. How do I run an OSMI node?**\ A complete, step-by-step guide on setting up and running an OSMI node can be found [here](/nodes/install-and-run-a-node). Get started and join the decentralized AI movement today! # Purchasing an OSMI Node License From the [Node page](https://www.osmi.ai/nodes) on the OSMI website locate the "Buy" button. That button will take you to the [Node purchase page](https://nodes.osmi.ai/). Once on that page, connect your Ethereum wallet. In order to purchase a Node NFT you will need $OSMI (for the node purchase) and Ethereum for gas fees. If you do not have $OSMI, you can obtain some by swapping another token for it through [Uniswap](https://app.uniswap.org/explore/tokens/ethereum/0x5485a469fAEa1492191Cfce7528AB6E58135aA4d). Follow the instructions on the purchase page and sign the transaction. When successful and complete, you will see your new OSMI Node NFT in your wallet. It will also show up near the buy portion of the purchase page. Once purchased, your node will show up in a section above the sale widget. Next to your unique Node NFT number there will be a Link button. By clicking on this button you will be able to generate a unique key tied to your wallet. You will need to copy this code and enter it when installing your node client software as outlined in this section: [installing and running your node](/nodes/install-and-run-a-node). # Node Tiers | Tier | License Price | Total Licenses per Tier | | --------------- | ----------------- | ----------------------- | | 1 | 10000 | 100 | | ignore the rest | keeping the table | we just need values | | 3 | 700 | 358 | | 4 | 800 | 313 | | 5 | 900 | 278 | | 6 | 1000 | 250 | | 7 | 1100 | 228 | | 8 | 1200 | 209 | | 9 | 1300 | 193 | | 10 | 1400 | 179 | | 11 | 1500 | 167 | | 12 | 1600 | 157 | | 13 | 1700 | 148 | | 14 | 1800 | 139 | | 15 | 1900 | 132 | | 16 | 2000 | 125 | | 17 | 2100 | 120 | | 18 | 2200 | 114 | | 19 | 2300 | 109 | | 20 | 2400 | 105 | | 21 | 2500 | 100 | # Install and Run a Node There are currently two types of OSMI Nodes that are both executable with a node NFT license, storage and inference nodes. Memory nodes handle storage needs for the OSMI ecosystem and do not require high end computers. Inference nodes, on the other hand, have heavy GPU, CPU, and memory requirements and will likely be run by fewer participants of the OSMI network, thus receiving a greater portion of the OSMI rewards. In the future, daily inference token generation counts will be calculated and added into the node reward pool (i.e. "tokens for tokens"). ## OSMI Super Node With one node install package you can run any type of node. **The OSMI Super Node is your main entry point into the OSMI node ecosystem and the only OSMI software you will need to install**. By running this node software, you will be able to link your Node NFT and run any type of OSMI node. This super node will be able to determine what type of node workload you'll be able to run in the OSMI ecosystem, depending on your node's system specifications. ### Linux On Linux, you will need to enter these instructions using a terminal window: ``` # IMPORTANT: Keep the user service running when logged out loginctl enable-linger $USER # Install support software - you may need to use sudo apt update && apt install -y unzip build-essential curl wget git # Get and unzip osmi-node wget https://links.osmi.ai/osmi-node_linux_amd64.zip unzip osmi-node_linux_amd64.zip # Install the software - you will be asked for your pairing code from the website ./osmi-node install # When you get this prompt, use the code from the nodes webpage used to link # Please paste your activation code and press enter to continue: ``` For some systems you will need to open up or port forward a specific port for OSMI to run (this applies mostly to home ISPs). The OSMI super node needs to communicate with our servers through UDP/TCP on port 8670. ### Windows For Windows PCs, download this [installer](https://links.osmi.ai/windows/OsmiNodeInstaller.exe) and run it. This current installer is not code signed, so you may see a warning that this publisher is not identifiable. We will eventually get code-signing enabled, but you can run the installer by clicking on run anyway. After running the installer a terminal window will open asking for the sctivation code. This is the code you receive from the "Link" button in the node page. Cut and paste it into the window and you're all set! ## Memory Nodes Because of their lower specifications, memory node workloads will be more widely run and thus receive a smaller allocation of node rewards when more nodes come online (all rewards are divided equally among other memory nodes). Memory Nodes currently run on the Linux or Windows operating system and have the following system recommended requirements: * Ubuntu 22.04+ * 4 or more CPU cores * 6GB or more memory * at least 100GB of hard drive space\ Below are the command line instructions to install and run this node: ## LLM/Inference Nodes (Coming Soon) Inference nodes require much higher specifications than memory nodes and running them requires a more thorough setup process and custom verification software to ensure these nodes are properly performing. If interested, please contact us through our various [communication channels](/introduction/links). We will be updating this space with install instructions soon. ### Preliminary Specifications: Some preliminary specs for these nodes are as follows (**these are only guidelines and will change**): * Linux Ubuntu 22.04+ * Nvidia based GPU (3080Ti or higher performance with 12GB VRAM) * 16GB+ memory * x64 CPU with 6+ cores * 250GB+ disk space ### Inference Node Distributions Currently inference Nodes receive rewards in the same pool as memory nodes. Because they require much higher specifications and provide more compute power for the network, they will be in receiving distribution in their own pool and will be eligible for "token for tokens" $OSMI rewards. In the future, token metrics will be recorded. Tokens in AI are the actual inference results (e.g. words are made up of tokens). With OSMI's API and product offerings these tokens are generated by decentralized inference nodes and paid for by customers. When an inference node generates tokens it qualifies for the daily reward of those $OSMI tokens, thus "tokens for tokens". Also, we anticipate, fewer Node NFT owners will run systems powerful enough to do inference, so the daily reward distributions will be delivered to fewer Node NFTs. We anticipate this will be a self balancing ecosystem among all the node workload types due to computational needs. In the future, node software will automatically detect if a node meets the requirements to run inference and will serve the network as an inference node. # Node Token Distributions The native token of the network, $OSMI, is distributed daily to node owners in a transparent and fair manner. Before being distributed, the following steps are performed: 1. Node Verification 2. Distribution Calculation 3. Vesting 4. Token Distribution ## Verification Before the distribution, each node is verified to ensure it meets the necessary criteria to receive $OSMI tokens. These criteria include being an active participant in the network, performing tasks such as inference, or contributing to the network's ecosystem (e.g. file storage). A node with a running uptime of at least 18 hours is verified to receive the minimum distribution (we will be lenient on this requirement for the first few days of the initial sale). Nodes performing direct AI tasks, such as inferencing, will receive a larger portion of the distribution. ## Distribution Calculation The daily distribution of $OSMI tokens is calculated based on the number of verified nodes, the total daily allocation of tokens, and a set distribution ratio. The total number of tokens to be distributed per day and the distribution ratio can be seen in the [Tokenomics](/tokenomics) section. **Node distributions group rewards according to node type**. For example, memory nodes split the node rewards amongst all other memory nodes, whereas LLM nodes distribute their portion among other LLM nodes. Currently, **all nodes participate in the daily distribution as one node type**, but in the future there will be other nodes introduced and they will also split rewards. This ensures a healthy ecosystem with node owners will be able to switch and provide resources to balance out the network. ## Vesting Tokens undergo a vesting period prior to being released and minted. This vesting mechanism helps deter abuse and promotes a healthy network. Tokens transition through two states: Held and Released. Tokens in the Held state are in a vested condition and cannot be redeemed. Released tokens, on the other hand, are no longer part of the Held pool and can be minted and transferred. ### 90 Days Tokens are held, or vested, for a span of 90 days. During this time, they accumulate, but remain inaccessible for any interaction. For instance, when a node owner operates their node for 100 days, 90 days of distributed tokens will be in a held state, while 10 days of tokens will be in a released state, released one at a time (i.e. day 91, tokens from day 1 distribution move to released state, day 92, tokens from day 2 distribution move to released state, etc.). Once in released state tokens are ready for minting to the user's preferred blockchain (Ethereum or GalaChain). It is crucial to note that tokens are only released on a daily basis from their vesting status.

One Year of Vesting

### **Forfeit of Vested Tokens** **If at anytime, a Node NFT is transferred to another wallet, the vesting period resets and all held tokens are burned. Released tokens will be unaffected and can be released to the original wallet.** ### **Accelerated Vesting** In the near future, node owners will have the option to burn held tokens to accelerate their vesting schedule. We will use community feedback to help determine best practices; stay tuned. ## Token Minting According to the above vesting schedule, once tokens are released from being held, a node owner has the option of minting their tokens (potentially incurring fees depending on which blockchain they are minted on). Once minted these tokens behave as any crypto asset free from restriction. # Node NFT To become a Node owner, users will need to purchase a Node NFT:
Contract TypeERC-721
Contract Address0xbf9A19E7e926d5D4c1a789F76db0AF23cA9854AB
Contract Sourcehttps://github.com/osmiai/contracts
LockedYes, non-transferrable for 1 year
# Roadmap OSMI’s journey to decentralize AI has been marked by exciting milestones, from launching the initial PoC and node network in 2023 to the public beta release in 2024. With the foundation firmly in place, we’re ready to take OSMI to the next level with the $OSMI token launch in Q4 2024. This token will unlock core functionalities, allowing users to purchase node licenses, access AI inference capabilities, and utilize new image creation features, all while supporting a deflationary model with token burns. Looking ahead to 2025, OSMI will introduce staking rewards, a GPU marketplace for private data ingestion, and API access for commercial use, opening up decentralized AI capabilities to businesses and developers worldwide. Each step of our roadmap is designed to build a resilient, scalable, and community-driven ecosystem, making OSMI a trailblazer in the future of people-powered AI.

OSMI Roadmap

# Team POW: Use about webpage