We’re thrilled to announce the upcoming integration of Ritual - an open, sovereign execution layer for AI. In the coming weeks, Omnisea will join the Crypto x AI revolution by becoming the first protocol enabling the creation of aiNFT Drops powered by Ritual.

Motivation

We believe that in the future, a significant part of the content we consume will be created using artificial intelligence. This applies to Images, Video, Audio, 3D objects, and other works. At the same time, at Omnisea, we consistently pursue the vision of being The Everything Launchpad. In recent weeks, we have added support for ERC-404 and DN-404, as well as significantly improved video and audio Editions. At this point, recognizing the growing synergy between Crypto and AI, we want to become the leading protocol for creating AI Digital Assets.

aiNFT: powered by Ritual, integrated into Omnisea Launchpad

Ritual brings compute on-chain, enabling smart contract developers with a way to request computation to be done off-chain from Infernet Nodes and delivered to their consuming smart contracts.

Ritual supports classical and large language models (LLMs), including most models supported by scikit-learn as well as open-source models supported on HuggingFace.

Omnisea will give Creators a way to use any selected AI model (service) deployed on Ritual Infernet.

The Architecture of Omnisea x Ritual Drops

At first, we’ll aim to introduce aiNFTs to advanced users as well as produce some ready to free-mint example AI-powered Drops. With time, all of the technical details described below will be abstracted away.

Let’s quickly overview what the aiNFT creation will look like.

• At the Create Page, next to the existing Asset Types (ERC721Psi, ERC404, DN404), we’ll add the aiNFT option.

• Upon selection, the Creator will see a few new form fields:

  1. Model: select from the list of trusted AI models or type the name of your custom service deployed on Ritual Infernet. This service will be responsible for accepting your Minters’ requests (via Ritual Nodes) and AI inference (execution).

  2. Decoder (automatically set if the model was selected from the list. If not, type contract address for custom models - for very advanced users): a Ritual Node returns the result of AI computation to the NFT contract in raw bytes. This output is then passed to the selected Decoder contract. Use one of our premade Decoders or type your custom Decoder contract address. A Decoder is responsible for accepting a single bytes argument and returning the string that will be embedded into the Token Metadata (tokenURI) under the image or animation_url property. Omnisea’s NFT contract isn’t aware of the exact formatting of Ritual’s output (bytes) because each model (service) can return a different result. That’s why we allow to specify a Decoder that will have this exact knowledge (bytes decoding implementation), and therefore, Omnisea aiNFT contract will be able to handle every AI Model supported (deployed) by Ritual.

  3. Immutable / Mutable Metadata: If mutable, every minted NFT file can be recomputed again (even using a new prompt) after the token owner pays the ReInference Fee (used for OSEA Buybacks). If immutable, once an NFT successfully obtains its metadata (upon AI inference and Ritual Node delivery), it won’t be able to recompute it.

  4. (Optional) Permissioned Node: if specified, only a selected Ritual Node will be able to deliver the AI inference result to the aiNFT contract. It’s an additional security measure to ensure your metadata and AI execution integrity.

• On the aiNFT Drop Page, a Collector after clicking the “Mint” button, will be asked to type the prompt (the first iteration assumes the Text-to-Image, Text-to-Video, Text-to-Audio, Text-to-3d, etc.). This prompt will be sent along with the aiNFT → Ritual compute request.

• An NFT is minted without metadata.

• Ritual Node returns the AI inference result to the aiNFT contract, and (after decoding), it becomes a part of the newly set Token Metadata under the image or animation_url property.

• Ritual Node who delivered the result will receive the Inference Fee (in the native coin of a given chain, e.g. ETH for simplicity). Therefore, Omnisea will become a profitable and sustainable platform for Ritual Node runners.

• If, when creating, the Creator set the aiNFT collection as “Mutable”, any token owner can pay the ReInference Fee (also in the native coin) for another AI inference (metadata regeneration). This fee is equal to 1.5x of the initial Inference Fee and the exceeding 0.5x will be used strictly for OSEA Buybacks.

In the future, with a growing number of AI services already deployed on Ritual Infernet, and, therefore, a growing list of our predefined “trusted” models (selectable from the list), launching aiNFT should only take a few minutes even for beginners. By integrating aiNFTs into Omnisea, we aim not only to meet current market demands but also to revolutionize the way we interact with Artificial Intelligence Onchain.