not approved
Advanced Cardano NFT Search Engine
Current Project Status
Unfunded
Amount
Received
$0
Amount
Requested
$56,550
Percentage
Received
0.00%
Solution

Build a fast Cardano NFT search engine with faceted search functionality - enhanced by Argus’ image similarity (machine learning), and designed to accommodate future multi-chain support.

Problem

Cardano lacks any advanced, ecosystem-wide, NFT search and discovery tools. This makes NFT markets inefficient, limiting adoption and reducing market liquidity.

Impact / Alignment
Feasibility
Auditability

ArgusNFT

2 members

Advanced Cardano NFT Search Engine

Please describe your proposed solution.

<u>Building the most trusted, user-friendly NFT ecosystem on Cardano</u>

The overall NFT market has undergone incredible growth in recent years: Chainalysis estimates Ethereum NFT sales totaled $44.2 billion in 2021 – roughly 220 times the volume of sales in 2020. And this prodigious market growth has resulted in the creation of hundreds of millions of NFT assets being minted across multiple blockchains.

At the same time, this dramatic increase in market size has created a great deal of competition to attract NFT users: From NFT-focused, Layer-1 blockchains like Flow, to NFT-dedicated, Ethereum Layer-2 scaling systems like Immutable, to ever-innovating NFT standards, platforms and projects across the space – the explosive growth of the market has given rise to a Cambrian explosion of new use cases, more functionality and better UX.

Cardano is well-positioned to excel in the competition between NFT ecosystems: Its strengths include superior safety and security of its NFT marketplaces; a small carbon footprint and low transaction fees for minting NFTs; resistance to DDoS attacks; and Cardano’s commercial focus on the developing world, which will open up global marketplaces to countless creatives in the coming years.

And while the NFT sector is exploding with innovation, there are still obvious improvements to be made to the market landscape that will go a long way in fostering broader adoption of NFTs – improvements that can also help fuel Cardano’s growth. Our team at ArgusNFT is focused on bringing some of the most impactful of these improvements to Cardano first. We build innovative NFT market infrastructure, and we’re focused on two overarching goals:

1) Building Trust: As in broader crypto markets, there are too many NFT scams. And one of the most pervasive scams in the space is fake/copy NFTs and plagiarism. This not only harms collectors. The theft of art and other IP harms creators and closed their minds to the opportunities NFTs create, so we’re building tools to greatly reduce the ability of scammers to sell counterfeit or otherwise plagiarized NFTs.

2) Improving market liquidity: Currently, in order to navigate NFT ecosystems in a safe and circumspect way, you pretty much need to be a crypto enthusiast with a good deal of time. This is obviously an impediment to NFT adoption, so we want to make markets more accessible and liquid by making searching for NFTs easier and more customizable, and making the UX of browsing for NFTs more customized to user preferences.

<u>Bespoke Data Solution for an NFT Search Engine</u>

One of the features that’s most obviously lacking in NFT markets – not only on Cardano but across all permissionless blockchains – is optimized search functionality on marketplaces, and a low latency, high throughput search engine capable of performing advanced (i.e. highly customizable) searches of all NFTs across an ecosystem. While there are many, barely functional NFT “search engines,” and a few projects (which we’ll discuss below) promising to launch more performant NFT search engines in the future, there are no existing, highly performant search engines for NFTs on any permissionless blockchain or open marketplace we’re aware of. There are numerous platforms that allow users to first look up a collection by its name, and then filter some NFT collections by attributes. But direct searches of attributes or NFT characteristics yield no results, or results that are incomplete and/or irrelevant. This greatly limits the visibility of small collections on major marketplaces, and makes it so the only way to get a circumspect view of an NFT ecosystem is through a labyrinthine search across twitter pages and Discord servers. And this market inefficiency is major impediment for adoption by new creators and new collectors.

As experts in ecommerce software development, our team believes the poor performance exhibited by current NFT search engines is primarily caused by them employing inappropriate data storage models. While relational databases are relatively simple to construct and ideal for certain applications, they’re rigidity doesn’t lend itself well to mapping complex, conditional relationships between data, or result in flexible, relevant full-text searches. And they run into increasingly bad performance problems when large amounts of data are continually ingested into the database.

For this reason, our team will utilize a platform for data ingestion, storage, and retrieval that’s geared towards searches of human readable text. Our approach will allow for the ingestion of vast quantities of data from a number of different sources, and the quick retrieval of relevant data based on full-text searches. It’s affordable to implement on a small scale, which will allow us to refine indexing configurations, optimize search results, develop partnerships, and build up B2C revenue streams to keep the search engine sustainable as traffic and maintenance costs increase. And the approach is also extremely scalable; it’s the same one used by some of the world’s largest online retailers and search engines.

<u>V1 - Advanced Search Functionality</u>

(primary deliverable for this proposal)

The ArgusNFT search engine will enable highly customizable searches of Cardano NFTs. In version one of the engine, users will initiate a search by typing one or more search terms in the search bar. These words can include asset name, collection title, an approximation of either asset or collection title, asset ID, policy ID, and/or any combination of NFT traits/attributes. Findings will be retrieved and ranked primarily based on the presence and prominence of search terms in NFT metadata.

As with existing NFT search engines, once search terms are entered and queried, relevant NFTs will be displayed by asset names and thumbnail images. But in addition to this thumbnail/name display of relevant NFTs, the engine will display “faceted search” checklists to the left of the thumbnail findings. In version one of the engine the first checklist will enable users to filter results by “all NFTs” or “NFTs listed for sale.” The second checklist will enable users to filter findings based on NFT collections. And the third list will allow users to filter their search by the most relevant NFT traits/attributes to the search terms.

For example, a search for “bear with a red hat and sunglasses,” the search may yield a checklist where users can filter by five collections containing NFTs whose metadata contains text highly relevant to “bear, red hat, sunglasses” and another checklist of the 10 most common traits/attributes accompanying these keywords in NFT metadata. This faceted search feature will function much the same way it does on leading online marketplaces like Amazon, where checklists are displayed alongside search results that users can use to quickly and efficiently narrow their search by such traits as brand/manufacturer, price range, and specific features and attributes. And with subsequent improvements to the search engine, it will include additional, faceted search checklists like “price range” and “collection size” to further improve user experience.

<u>V2 - NFT Search Engine x Argus Machine Learning</u>

(not within the scope of this proposal)

At the heart of ArgusNFT’s product and service suite is a custom, Machine Learning model for determining and quantifying the similarity of NFT images. Our Catalyst proposal, Using this model to detect fake/copy NFTs has been our team’s primary motivation for developing this ML model. And just days ago our team received major market confirmation of the long-term demand for this solution when we finalized a partnership with JPG Store - Cardano's leading NFT marketplace - to protect the Store's inventory from counterfeit NFTs. But the ability to quantify image similarity between NFTs is not just useful for spotting "fake" NFTs; it can also bring huge value to NFT search engine user experience (UX).

For example, imagine you’re browsing through a collection on an NFT marketplace and you love a particular NFT, but it’s just too expensive, or it’s not for sale, so you want to find the NFTs in the collection that are most similar to it. Currently, on NFT Marketplaces, you’d have to click on the NFT you like, write down its attributes, then start checking and unchecking faceted search checklists for who-knows-how-long, trying to find NFTs for sale with similar attributes. But once Argus’ ML protocol is refined and integrated into the ArgusNFT search engine, finding the most similar NFT to one you love will be two simple clicks: You’ll just need to click on the NFT you love, then click “Find Similar NFTs,” and the most similar NFTs will be displayed, ranked by similarity according to Argus.

Another example – imagine you’re on twitter on your mobile phone, and you see profile pic you love, of what you think is a Cardano NFT, but you know nothing else about the image. Currently, you could ask the person with the NFT about it, or depending on the nature of the image, you may be able to search on Twitter or Google for the collection it belongs to. But with Argus’ ML you can simply take a screenshot of it or save it, run it through the Argus NFT search engine, and if it’s a Cardano NFT it will be delivered as a search result – along with the collection it belongs to and where it’s being held or listed for sale.

For a last example – imagine you’re patiently trying to find the right NFT(s) in a collection (or collections) you really like. But you don’t have a ton of time to spend searching, and for this reason you search less frequently and keep missing out on the best deals. Well, by virtue of the Argus ML model, if you search repeatedly on the ArgusNFT search engine and connect your wallet when you do, the search engine will show you suggested NFTs based on your past searches, click-throughs and/or purchases, saving you time and alerting you to NFTs you might have otherwise missed.

Please describe how your proposed solution will address the Challenge that you have submitted it in.

The current lack of performant NFT discovery tools leads to a fragmented, highly inefficient market. Limited search function on marketplaces leads to poor user experience, and marketplaces are actually incentivized to conceal the location of NFTs that are listed on other marketplaces. The lack of a single place to see multiple marketplace inventories makes it near impossible for new marketplaces - even innovative ones - to gain much market share. And suboptimal user experience equates to a suboptimal level of adoption from NFT collectors accustomed to other blockchains - some of which have marketplace aggregators and cross-chain marketplaces.

All that being said, a performant search engine with faceted search will improve user experience and increase NFT liquidity by making listed NFTs more easily discoverable to potential buyers. Additional search features based on image similarity machine learning will further improve UX and NFT discovery. And while it’s not part of the scope of this proposal, the multichain support we intend to integrate on the search engine in the future will act as a window into the Cardano NFT ecosystem to NFT collectors accustomed to trading on other blockchains.

What are the main risks that could prevent you from delivering the project successfully and please explain how you will mitigate each risk?

While our team has years of experience building and optimizing search engines for web2 applications, there are differences in the nature of blockchain data and the data inputs of web2 search engines that will present unique challenges to optimizing an NFT search engine on Cardano. And while metadata formatting standards have been widely adopted in the ecosystem, the accuracy of the ArgusNFT search engine will correspond to the appropriateness of NFT metadata for processing by a search engine.

That said, we’re confident we can produce a far better NFT search engine than others that have launched. Most NFT Search engines that are already online have little to no practical utility because they’re either too slow, support too little of a blockchain ecosystem, or both. There are a few NFT search engines that do have limited utility, including <https://polygon-nfts.com/>, which provides reasonably responsive and accurate search of some of Polygon NFTs, and <https://nftsearch.site/#>!, by Numbers Protocol, which identifies whether images uploaded to it are Ethereum NFTs, and where those NFTs. However, even these examples have some serious shortcomings:

  • They don’t have faceted search features, which severely limits their utility.
  • They only support a fraction of the NFTs in the ecosystems they claim to support.
  • They don’t have any image similarity component to rank search results by similarity to a given NFT.
  • They don’t support Cardano, nor are they planning to support Cardano in their roadmap.

That being said, the more significant, potential competitors of the ArgusNFT search engine are NFT search engines that have yet to launch. These include Quoth, NOOFT, Ludo and HEBYS. However, none of these search engines is building on (or planning to build on) Cardano, which is justification in itself for a qualified team to be working on an NFT search engine on Cardano - so Cardano NFT users maintain access to comparable or superior tools as NFT users on other leading blockchains. But without exception, these projects are also very thin on explanations or documentation about what they’re building. And none of them, to our knowledge, will integrate image similarity machine learning into their engine to enhance its functionality and UX.

An additional risk is that even if we’re first to market with a highly performant, NFT search engine, the Cardano ecosystem won’t use it. But while it’s true that some Cardano NFT collectors are perfectly fine with the current market landscape, we believe newer entrants and collectors who want an edge in having a more circumspect view of the market will find the ArgusNFT search engine to be the most efficient way to search the entire ecosystem.

That said, we know that marketing is important, especially in the period immediately following the platform’s launch or major upgrade, so we’re budgeting some funds for a social media marketing budget to help spread the word during those critical periods.

Likewise, developing partnerships with Cardano marketplaces for them to use the search engine API, and partnerships with NFT projects and other platforms for them to advertise via the search engine will be critical for the projects long term sustainability and independence from continued Catalyst funding, so we've budgeted for an initial push in these business development initiatives.

Please provide a detailed plan, including timeline and key milestones for delivering your proposal.

This proposal is for the development of Version 1 of the search engine, which will enable users to run faceted search queries based on NFT data and metadata.

The plan is to leverage Cardano DB Sync (an Open Source component developed by IOG) to extract all Cardano NFTs and their metadata and build an ElasticSearch index that will be used to provide search functionalities.

NFT data and Metadata will have to be parsed, processed and indexed in order to provide powerful query capabilities. An example could be represented by traits contained in a SpaceBud's metadata such as the following one (pool.pm link)

<https://pool.pm/d5e6bf0500378d4f0da4e8dde6becec7621cd8cbf5cbb9b87013d4cc.SpaceBud1576>)

Traits

  • Chestplate
  • Belt
  • X-Ray
  • Jo-Jo
  • Type
  • Bull

Additionally, if any NFTs are listed on any Cardano marketplace, we will index additional information such as price, collection floor price, etc.

Tasks and Deliverables

UI and UX

Search Page Components

  • Card for NFT
  • Search bar
  • Advanced Search
  • Facets
  • Pagination

Ops and Engineering

  • Extract NFT metadata (1 month) - September
  1. Create Cardano DB Sync integration component to extract list of Cardano NFTs
  2. NFT Metadata Parser for multiple standards
  • ElasticSearch indexing (1 month) - October
  1. Scheduled cron job to update NFTs index: create new records, delete burned NFTs, update re-minted NFTs.
  2. Search Results
  3. Policy ID
  4. Assent name
  5. Fingerprint
  6. Asset ID
  7. Metadata
  8. Argus Integration: flag potential fakes
  9. If NFT belongs to Verified collection
  10. Artist if available
  11. Twitter account if available
  12. Project website details if available
  13. If NFT for sale
  14. Marketplace on which NFT is for sale
  15. Price if listed on marketplaces (best effort)
  • Fetch Marketplaces information (2 weeks) - October
  1. Find Script Addresses of Cardano NFT Marketplaces
  2. Extract prices for listed NFTs
  3. Index Listed NFTs on Cardano Marketplaces
  4. Update index to reflect delisted/sold NFTs
  • ElasticSearch Operation (2 weeks) - November
  1. Configure Elasticsearch on Kubernetes
  • Landing Page (1 month) - November
  1. Setup page skeleton
  2. Search bar
  3. Results
  4. Sorting
  5. Facets: search refinement
  • API (2 month) December - January
  1. Search by policy
  2. Search by policy + asset
  3. Search by fingerprint
  4. Search by asset id
  5. Free text search (any or all of)
  6. Name option
  7. Body option
  8. Facets option
  9. Argus Integration (exclude potential fakes)
  10. Search base

Please provide a detailed budget breakdown.

UX/UI $1,800

Extract NFT Metadata - 1 Engineer Full Time 1 month. $10,000

ElasticSearch Indexing - 1 Engineer Full Time 1 month. $10,000

Configure, deploy and manage ElasticSearch 3 weeks. $7,500

Landing Page - 1 Engineer Full Time 1 month. $10,000

Search API - 1 Engineer Full Time 6 weeks. $15,000

Cloud Costs - ($500/month for 3 months). $1,500

Create and conduct Post-Launch Product Survey - $750

Social Media Manager - (part-time $600/month for 3 months) $1,800

Proofspace Subtotal $56,550

Please provide details of the people who will work on the project.

Argus NFT

ArgusNFT is developing a suite of cross-chain solutions to improve trust and discovery in in NFT markets. Argus, which is ArgusNFT’s flagship product, is an industry-first solution to the problem of fake/counterfeit NFTs and plagiarized artwork being sold as NFTs. It employs artificial intelligence and machine learning (AI/ML) to verify the minting history of NFTs, providing stakeholders information they need to determine whether an NFT is an original or a copy. Agrus currently supports all Cardano NFTs, and R&D is ongoing to integrate support of additional blockchains to protect against cross-chain copies. LinkedIn | Twitter

The members of the ArgusNFT team supporting this proposal include:

  • Giovanni Gargiulo - Cofounder of ArgusNFT. Principal Software Engineer with 15+ years of commercial experience. DevOps advocate with strong experience in Machine Learning, distributed systems and cloud. Cardano Plutus Pioneer, Stake Pool Operator and Open Source Developer. Linkedin, Git
  • Anthony Capitan - Cofounder of ArgusNFT. Entrepreneur with 7 years business development experience, 15 years sales, marketing and project management experience. Business development advisor to a number of early stage Cardano projects. Linkedin
  • Jaime Caso - Full Stack and Mobile Engineer - Stake Pool Operator and Cardano Wallet Developer. Linkedin, Git
  • Kurt Hartmann - Kurt Hartmann has 20+ years of experience as a developer and architect working as an IT consultant for companies in the financial, health, and retail industries. His areas of expertise include modernizing and integrating legacy systems, ETL batch jobs, building API data sources, and implementing workflow patterns using micro-services and messaging. He also has a solid background in database design and web services. He participated in the Cardano Plutus Pioneer Program (cohort #1) and has spent significant time learning the Haskell programming language. Git
  • Natasha Younge - received an MBA in management, and has provided social media consulting and management to several Web3 projects, where she focuses on Twitter and Twitter Spaces to cultivate community growth and engagement. LinkedIn

If you are funded, will you return to Catalyst in a later round for further funding? Please explain why / why not.

Depending on the availability of alternative funding sources, once this proposal is complete we'll possibly propose that project Catalyst fund the search engine's upgrade to V2 - integration of Argus' image similarity (machine learning) model in a future fund. This would significantly expand the search engine's functionality. However, in the mid-to-long term we have a plan to make the search engine financially self sustaining via multiple potential revenue streams - Ad revenue, sponsored links, fee for clicks converting to sales, and others.

Please describe what you will measure to track your project's progress, and how will you measure these?

  • Publication of project deliverables through monthly reports, which will demonstrate adherence to the work plan.
  • Post-launch, create, conduct and publish findings from a product survey of no less than 50 users to gauge user interest in the product, it's strengths and where there's room for improvement.
  • Post-launch, publish anonymized data from meetings with no less than three NFT marketplaces or other platforms - recording their interest in purchasing one of the services intended to make the platform financially sustainable: Ad revenue, sponsored links, fee for clicks converting to sales, etc.

What does success for this project look like?

We'll consider this project successful once:

1) We launch a performant search engine on Cardano;

2) we regularly market the search engine on social media and within our team's networks for roughly 90 days;

3) and we conduct a survey to gauge user interest in the product, it's strengths and where there's room for improvement.

Please provide information on whether this proposal is a continuation of a previously funded project in Catalyst or an entirely new one.

This is a new proposal.

close

Playlist

  • EP2: epoch_length

    Authored by: Darlington Kofa

    3m 24s
    Darlington Kofa
  • EP1: 'd' parameter

    Authored by: Darlington Kofa

    4m 3s
    Darlington Kofa
  • EP3: key_deposit

    Authored by: Darlington Kofa

    3m 48s
    Darlington Kofa
  • EP4: epoch_no

    Authored by: Darlington Kofa

    2m 16s
    Darlington Kofa
  • EP5: max_block_size

    Authored by: Darlington Kofa

    3m 14s
    Darlington Kofa
  • EP6: pool_deposit

    Authored by: Darlington Kofa

    3m 19s
    Darlington Kofa
  • EP7: max_tx_size

    Authored by: Darlington Kofa

    4m 59s
    Darlington Kofa
0:00
/
~0:00