Proposal to create an SNS DAO for ELNA



Update : Proposal is live

This is a follow-up post on the previous one we had shared about ELNA’s upcoming SNS sale.

> Upcoming ELNA SNS Decentralization Sale

The decentralization sale is scheduled to start shortly; ELNA SNS DAO is getting ready to debut.

PFA proposed ELNA SNS configuration > SNS/sns_init.yaml at main · elna-ai/SNS · GitHub

Upon acceptance of this proposal, the swap will automatically be started and be open for participation.

AI on Blockchain

AI democratization is the goal of ELNA, being the first fully on-chain generative AI platform in the world, producing AI agents that work flawlessly across all external platforms.

ELNA’s Main Offerings

  • Easy to use AI agent creation platform
  • A marketplace to monetize your Ai agents
  • A developer ecosystem for collaboration and listing-cum-monetization of your AI tools

What’s Inside ELNA

  • ELNA’s vector database implementation in canisters and Retrieval Augmented Generation (RAG) allows for faster data storage and retrieval
  • ELNA’s Prompt Engine customizes the AI agent’s behaviour precisely to your liking.
  • Connect any large language model inference engines to the ELNA platform for agent creation
  • Integration tools for your AI agent to interact with any platform - your website, X, LinkedIn, Instagram etc.

Why Participate?

  • Governance: ELNA token holders will play a crucial role in shaping the future of, from decision-making on project developments to the strategic direction of the platform.
  • Innovation: By participating, you support the integration of LLM Agents with blockchain technology, fostering innovation in how decentralized applications are built and interacted with.
  • Privacy & Transparency: is committed to ensuring user privacy and operational transparency, leveraging the inherent benefits of the ICP blockchain.

How to Support & Participate

You can support this vision of decentralized intelligence by participating in ELNA’s decentralization swap, for which we encourage you to read our Whitepaper and Tokenomics Sheet, as well as watch the DApp Walkthrough and Presentation Deck Video.

For consistent updates and enquiries -
Join our community on X & Discord

Your participation and support in the SNS decentralization sale not only contribute to the realization of an open AGI on the blockchain, but also mark your stake in the future of decentralized, transparent, and user-centric AI technologies.

ELNA’s Partnerships, Collabs & Featurings:

IC Ecosystem Partnerships/Collabs

AI agent Use Case Demo

Featured in

YouTube Lives / X Spaces with Community Leaders



NB : This isn’t financial advice, kindly invest post thorough research.


Dear ELNA Team,

I am deeply interested in the innovative approach ELNA is taking towards democratizing AI through blockchain technology, specifically on the Internet Computer platform. Your vision for leveraging the blockchain to create, monetize, and deploy AI agents is truly pioneering. However, I have a few technical inquiries to better understand the feasibility and execution of such an ambitious project.

  1. Model Training Infrastructure: Given the substantial data and computational power requirements for training high-quality AI models, could you elaborate on the infrastructure ELNA plans to use for this purpose? Traditional setups involve clusters with thousands of nodes, often utilizing GPU chips for efficiency. How does ELNA intend to navigate these requirements within the DFinity ecosystem, which, to my knowledge, currently lacks direct support for such configurations?
  2. User Request Processing: For AI models to function effectively and deliver rapid responses, they typically run on servers equipped with specialized hardware, such as GPUs. Could you provide details on the server infrastructure ELNA will employ to process user requests and ensure timely and accurate AI agent responses?
  3. Achievements and Benchmarks: I am curious about ELNA’s current achievements in terms of data volume for experiments, the types of servers used for these tests, the size of models successfully deployed (especially if loaded into canisters), and any performance metrics, such as token speed per second, that have been attained. This information would greatly aid in understanding the practical aspects of your implementation strategy.

Your project represents a significant step forward in the intersection of AI and blockchain technology. I appreciate any insights you can share regarding these technical aspects, as they are crucial for grasping the full scope and potential of ELNA’s capabilities.

Thank you for your time and consideration.

Best regards,


I have further questions that delve into the specifics of your model development and deployment strategy. This inquiry aims to build upon the insights you’ve already provided and seeks additional clarification on several key points.

  1. Adopting vs. Developing AI Models: Given the significant investments by companies like Google, Facebook, and OpenAI in their AI models, I am curious about ELNA’s specific approach. Will ELNA adapt these existing models or some other opensourced to address their limitations, or is there an intention to independently develop proprietary models? If the latter, what datasets and hardware infrastructure are envisioned to support this endeavor?
  2. Training Data and Computational Resources: Could you elaborate on the types of data ELNA plans to utilize for training its models and the expected hardware requirements? This information is crucial for understanding the scale and feasibility of your model training processes.
  3. Model Quality and Benchmarking Expectations: Are there any benchmarks or comparative analyses planned to evaluate ELNA’s models against those developed by the major tech giants mentioned earlier? Insights into your expectations for model quality and performance metrics would be highly valuable.

Your pioneering efforts at the intersection of AI and blockchain technology hold the potential to redefine how AI applications are developed and deployed. Further details on these technical and strategic aspects would greatly enhance our understanding of ELNA’s capabilities and its position within the broader AI ecosystem.

Looking forward to your detailed response.

Best regards,


Dfinity is the one that said they are working on GPU subnets so I would imagine this is part of their plans. Are you saying you’re not aware of Dfinity plans on having GPU subnets at some point?

It’s not yet clear when, what specifications, and how many. If ELNA is dependent on these nodes, it must be clearly stated that the project will wait until these nodes join the subnet. Before that, there’s a risk that ELNA will not fully operate on DFINITY. Moreover, it’s important to consider how many of such specific nodes would be needed to train models and process requests with adequate quality and speed.


Thank you @Alexander for the interest and your valuable inputs

To make it simple - ELNA is a Ai agent creation platform.
An agent is a combination of intelligence (LLM) + Knowledge (vector DB) and interaction (integration tools).
Please have a look at our video

From our Architectural block you can see that we are we are creating a creator cum developer ecosystem. Where developers could list their tools, llms etc … and creators could create agents with use of these available tools. Finally we have a marketplace and external integration.

We help to integrate, deploy and monetise.

Along with orchestration part we also create on chain components for AI for example vector db, prompt engine, inference etc …

Now coming to your questions
We need GPU canisters to run a larger model on chain and do fine tuning. Currently without GPU only smaller models can be run on chain (also there is instruction limit and memory limits) we also support off chain LLM integration for agent creation.
Ones the GPU canisters are live we could support larger models to run completely on chain.

Our approach is creating a community driven ecosystem that could bridge the gap between development and application creation.

1 Like

I like the two token system… using ELNA tokens to purchase Elixr tokens creates that deflationary token aspect I look for - I believe you’re burning the ELNA used to purchase Elixr making ELNA deflationary from day 1, can you please confirm my assumption?

Many folks had the foresight to participate in the LBP via the Sonic platform last year. Are those tokens that were sent to participants of the LBP able to be used in staking post a successful SNS launch or will new tokens have to be issued to replace them at some point in the future?

I hope you can grow into a community platform for AI creators i.e. like a better version of SingularityNet. Being on the IC will give you a leg up. I look forward to your launch proposal.

1 Like

Yes @DHemingway ,
you are absolutely right.

You need to burn ELNA tokens to purchase Elixr tokens, which is used in all the in-app purchased inside ELNA platform including creation and usage of agents.

Our core vision is to create an AI ecosystem by the community for everyone.

When are you guys planning to launch SNS and of course i need your token paper.

1 Like

Hi @Antony
The proposal is live : NNS Dapp


There is no comprehensive emc, but emc was rejected, and now almost no one has asked about the new sns project.