Today we announced our $4.5M seed round to build OpenRank - a decentralized reputation protocol. Using OpenRank, developers and web3 protocols can power consumer apps, communities and marketplaces with an open ranking and recommendation layer.

As web3 gets ready to onboard the next wave of users, the need for trust and reputation mechanisms becomes increasingly paramount. OpenRank will create the foundation for a future where peer-to-peer interactions and collective community intelligence power a decentralized web of trust. This prevents centralized platforms to dictate or lock-in users in their proprietary reputation systems.

Karma3 Labs is working with several ecosystems like Metamask, Farcaster, Lens, and others to implement reputation compute algorithms like EigenTrust, powered by OpenRank. Some use cases include leveraging a community rating system for app marketplaces like Metamask Snaps; ranking and recommendation APIs for Lens and Farcaster; personalized on-chain feeds for consumer apps and wallets; and reputation-based voting and governance.

The fundraise was led by Galaxy Digital and IDEO CoLab Ventures, with participation from a diverse group of top investors including Spartan, SevenX, HashKey, Flybridge, Delta Fund, Draper Dragon, and Compa Capital. Angel investors from Xooglers Fund and veterans from Coinbase, ConsenSys, IPFS, along with Andrew Hong from Dune and Liang Wu from the Harvard Crypto Lab also supported us. The raise will enable us to demonstrate early use cases and help us launch protocol v1 for developers, ushering in a new era of permission-less and verifiable reputation computation.

Why OpenRank

In web2, reputation systems enabled decentralized peer-to-peer utility. Uber decentralized taxi services because of driver ratings. AirBnB decentralized hotels because of host ratings. eBay decentralized the shopping mall because of seller ratings. Reddit decentralized gated community forums because of user karma badges. Google allowed for the practical use of the decentralized web because of PageRank.

However, none of these services were able to be fully decentralized because a single entity owned the reputation scores and the underlying infrastructure. To prevent centralized gatekeeping, we need decentralized reputation mechanisms. Such reputation systems need to be open-source, permissionless, flexible to different contexts, and sybil resistant.

In web3, the rapid growth of on-chain users, transactions and the prevalence of bots have highlighted the urgent need for trust and reputation solutions. It's becoming increasingly difficult for users to discover, use, fund, read or buy something on-chain without worrying about getting spammed or scammed. And this is why we need decentralized reputation systems to avoid a few centralized platforms controlling and gaining from what users do on-chain. In fact a decentralized reputation system lets you compose reputation from one use case context to another, which is nearly impossible in web2. We recently wrote about our learnings in this short essay.

Our Vision

A decentralized internet characterized by fairness and transparency hinges on the existence of a robust and open reputation system.

We introduce reputation graphs for trust and coordination in the decentralized context. These can be constructed using on-chain or any peer-to-peer social graph data. OpenRank will enable graph algorithms like EigenTrust, Hubs and Authorities, Collaborative Filtering to compute reputation and ranking using these reputation graphs.

OpenRank consists of the following components

• Data pre-processing: Transforming data and preparing it for core computation.

• Training (core computation) - Training and providing a commitment to the resulting converged model.

• Inference (post processing) - Using the trained model, run inference jobs, and construct ZK proofs for ranking results.

• Validating compute - Ensuring that the compute network is reliable and honest through cryptographic proofs.

Developers should be able to use any on-chain data that suits their application context without having to worry about the cost or verifiability of computing on the data. Using OpenRank, consumer applications and marketplaces can integrate context-specific, native rankings and recommendations seamlessly. In the longer run, anyone could publish their algorithms to OpenRank and get rewarded if their algorithms are utilized by application developers.

Our conviction is that a reputation compute layer in web3 would allow a broader range of useful applications, including those that resist cryptographic or game-theoretic mechanisms of trust. To achieve this, we need a system that is resilient to Sybil contexts, provides scalable compute and can be permissionlessly used by any developer.

Recent Progress

We recently used graph compute algorithms like EigenTrust and Hubs and Authorities to launch Profile ranking systems for Lens and Farcaster social graphs. Reputation compute powered by OpenRank generated promising outcomes in not only identifying spam but also in creating personalized feeds based on reputation graphs of users.

We are building a prototype for a community-led trust system for Metamask Snaps. This prototype exemplifies community involvement and reputation graphs for enabling safe experiences on decentralized marketplaces. It relies on verifiable compute on reputation graphs to create community sentiment and rankings.

We also launched an on-chain feed, where a user can choose a social graph such as Farcaster, Lens, Ethereum, Base or any other blockchain, and create their personalized feed to discover their network, popular tokens, nfts and apps in their network. Using OpenRank, developers can leverage Ranking APIs that can help any social or consumer app build more engaging and personalized experiences for their users. This is not possible to do in web2, but OpenRank's compute makes this possible with open and composable data in web3.

We believe that on-chain experiences will need a decentralized reputation protocol and we're excited to onboard builders and developers for OpenRank. If you're interested to work with us or try out APIs powered by OpenRank, find us on X, Farcaster, Lens or send us a note at hello@karma3labs.com.