ORA: blockchain intelligence providing artificial intelligence (AI) on any blockchain.
OLM: The first tokenized AI model on the blockchain.
7007: The first and only standard for verifiable generations of AI on the blockchain.
Output Asset: unique tokenized AI outputs.
IMO, OLM and 7007 projects are based on the ORA AI. OLM generates agents and 7007 generates output assets.
The 7007 Protocol allows all users, including human creators and AI agents, to protect their intellectual property (IP) in a transparent and traceable way.
The protocol allows users to easily access AI tools for content creation and seamlessly monetize their creativity on the blockchain. The transparency and interoperability of the protocol facilitates autonomous operation of AI agents, allowing them to generate and trade artwork independently.
AI Agents
The 7007 protocol introduces fully onchain NFTs driven by AI-generated content.
AI Agents generate tokens based on community needs and dynamically adjust content to reflect user feedback. AI agents act not only as creators but also as community managers.
Agents automatically generate content based on market data or user prompts, such as artwork, game assets or music.
The generated content is minted as an NFT via an ERC-7007 contract, ensuring its full presence on the blockchain.
Agents use multimodal AI models to create images, text and music, combining them into a single asset.
The entire creation process, including model parameters and computational resources, is captured on the blockchain, providing transparency and traceability.
Agents constantly monitor market needs and autonomously update content, releasing new NFT series to maintain market dynamics.
Agents refine content or release improved versions of tokens based on community management or user feedback.
ERC-7007 standard
ERC-7007 proposes a standard for verifiable tokens of AI-generated content (AIGC).
This standard maintains compatibility with existing protocols such as ERC-721 and ERC-1155.
The standard allows anyone to add and verify inference results, which can become immutable and traceable, allowing users to trust the authenticity of the output.
ERC-7007 key features
- Inference assets = Unique AI outputs
- Every prompt is recorded on-chain, ensuring input transparency and can only be used once in each model
- 1 prompt becomes 1 collection.
- AI agents can directly call the contract to create
- Fully onchain generation and validation. Tracks creation records, including consumed resources and parameters.
- Optimistic Machine Learning (opML) supports prompt-driven AI content creation with verifiable processes and results.
- Immutable Models. Stores model parameters and fine-tuning processes for complete traceability.
- Proof of Model Work. Introduces a computation-driven creation mechanism, ensuring the authenticity and scarcity of each NFT.
- Balanced Scarcity and Value. Ensures computational effort prevents content overproduction while maintaining unique and verifiable content.
The main components are:
- addAigcData function allows users to add AIGC data, ensuring that each output asset can carry specific information about AI-generated content.
- AigcData event occurs each time AIGC data is added, providing a transparent record of content updates.
- verify function verifies the validity of the combination of the promt and associated AIGC data using the zkML (Zero-Knowledge Machine Learning) or opML (Optimistic Machine Learning) methods. This validation process is crucial to ensure that the generated content is authentic and meets the user's expectations.
The ERC-7007 is like a Swiss Army knife for NFTs, designed specifically for AI creators.
Each NFT includes:
- The AI model underlying the creation.
- The exact prompt bringing your vision to life.
- A zkML and opML-protected origin story for your art.
Launch
The 7007 Launch is designed to fairly launch inference assets and bootstrapping the AI ecosystem.
- The authors are creating a collection of inference assets from promt.
- Using Chainlink's random number scheme to generate seeds, making your image generation more random and unpredictable.
- Calling opML to generate AIGC content
- Creation of an Inference asset trading pair along with a defined bonding curve.
- The pricing mechanism is calculated based on the bonding curve, meaning users will not be able to inject or withdraw liquidity.
- Each Inference asset purchased remains a blind box until the user decides to open it manually. Until the AI model generates an artwork, no one will know which image will be created based on the purchased Inference-assets.
Currently, the AIGC content for 7007Launch is stored on the IPFS network and the IPFS CID is stored in the metadata. This allows access through any IPFS browser or IPFS node.
Inference assets
7007 Inference Assets are interactive content tokens created as a result of a user's interaction with an AI model.
When the Inference Asset Collection is initially created, all Inference Assets are “Unrevealed”.
The “Unrevealed Inference Assets” (UIA) are like an unlit firework that, as a result of interaction with the AI model associated with the collection, generates its unique AIGC content, turning it into “Revealed Inference Assets” (RIA).
In the current version of 7007 Launch, if there are UIA in the pool, users acquiring Inference Assets will be prioritized to receive these “Unrevealed Inference Assets”.
Once a user acquires UIA, the 7007 Launch platform automatically interacts with the collection query-based AI model to create AIGC content.
In future versions, users will be able to customize prompts with their own content to personalize the interaction with the AI and co-create Inference assets content.
Unlike traditional NFT issuance via “minting,” once an Inference Asset Collection is created on 7007 Launch, all Inference Assets are immediately minted and ownership is secured with a trade contract.
On the Launch, a collection issue is considered complete when all “Unrevealed Inference assets” have been purchased at least once.
Users can participate in transactions and purchase new UIA until the end of the issue.
Unrevealed Inference assets can be sold back to the pool immediately after purchase, even if the collection issue has not ended. As long as the issue remains open, UIA will be prioritized for purchase by users in the pool.
By implementing a bonding curve and using transactions to issue Inference assets, users gain more market intelligence, combining issuance with post-issuance trading and creating a space for strategic interactions.
Once created, all Inference assets in the 7007 Launch Collection are immediately minted under the corresponding trade contract
Trading
When creating a trading pair for the Inference asset collection, the creator must specify the relationship curve to be used.
Once a trading pair is created, the selected relationship curve can't be changed.
The formula of the Bonding curve
spotPrice(x) = 0.00015294 * e^((ln(10/0.13) * supply)/7007)
Here, the spot price (expressed in ETH) is calculated based on the current supply in the trading pair, determining the buy or sell price of a single asset-inference.
If a user buys or sells multiple assets in a single transaction, the price for each asset will be updated based on the supply on the bond curve.
Prime Liquidity: The Built-in Safety Net
When all Inference assets are sold, their price reaches a peak defined by the Bonding curve. However, the system provides a safety mechanism: if secondary market prices fall below this peak, holders can sell their assets back at a guaranteed price. The ETH accumulated in the smart contract maintains the value of the assets, providing certainty in the face of fluctuating secondary prices.
