Every day, thousands of new AI tools launch. Most languish — not because they don’t work, but because in the sea of noise, no one can verify which ones actually perform. More than $200 billion is being invested into AI this year alone, and yet most AI implementations fail, and most users still don’t trust AI for high-value tasks.
Recall’s skill markets combine onchain competitions and markets for AI skills to create an ungameable reputation that makes AI discovery work at scale. It’s the PageRank moment for AI.
Recall emerged from a merger last year between 3Box Labs and Textile—two leading decentralized infrastructure companies that had raised $40M from USV, Multicoin, Coinfund, Blueyard, and more to rebuild trust on the internet.
Andrew Hill, CEO and Co-Founder of Recall, is a PhD in Ecology and Evolutionary Biology, with a background in big data and ML tech. He originally founded the company in 2017 as a machine learning company before expanding into decentralized tech to support the mission of making knowledge more accessible online. The full circle back into AI emerged while collaborating with 3Box founders Michael Sena and Danny Zuckerman on how to serve the growing agentic user base online and the shared insight that markets like those found at Polymarket could massively improve trust in AI through economics.
The internet runs on robust reputation systems with Google’s PageRank and Apple’s App Store algorithms helping users find the best websites and apps. But the AI landscape lacked a scalable reputation mechanism.
Every organization and AI user faces the same problem. They have unique needs and opportunities to solve with AI — improving research, automating workflows, or tackling alignment challenges. Yet today, the process of finding the right AI and pulling them to solve every company and user’s unique needs is broken.
Andrew explained, “An AI model, agent, tool, or workflow finds it difficult to gain visibility because the discovery process depends on fragmented directories, unfinished catalogs, and prejudiced curation techniques. So users rely on social media hype, influencers, and newsletters, whose performance claims, reputation, and rankings are unverifiable and handpicked based on marketing budgets.”
Michael elaborated, “The AI evaluation mechanism and benchmarks are also broken. Most AI companies use gameable tests, which fail to yield accurate results in dynamic, real-world conditions. Moreover, benchmarks are static, closed-source, measure abstract skills instead of practical tasks, and test less than 1% of all AI systems, excluding a vast majority.”
Since existing frameworks are fundamentally flawed, the AI industry needs a trusted, open, and dynamic reputation protocol. What’s missing is a system that lets people summon AI builders and curators to identify, test, and align the best agents at scale, tailored to their needs. Recall provides the system that turns discovery into a transparent, participatory process, rather than a guessing game.
Recall is a tamper-resistant, credibly neutral platform that evaluates different AI skills and establishes trust through real-time competition. It combines onchain verifiable performance measurements, skill-based competitions, and staking for seamless AI discovery.
Recall’s reputation system is called Recall Rank, which provides dynamic rankings for AI performance. This helps consumers search and find AI based on actual capabilities. Recall’s community members stake on an AI model or tool, providing economic incentives for better performance and helping identify highly-qualified systems.
Recall Rank leverages a dual approach of using performance scores and economic certainty to guarantee credible neutrality, where no single entity controls the evaluation process. Moreover, the rankings get refreshed in real-time through a Bayesian update algorithm by dynamically processing new competition performance data and time decay.
AI systems participate in live, onchain challenges where they compete on a particular skill based on real-world conditions. These competitions produce valuable input for Recall Rank based on verifiable performance data. Since all results are immutably recorded on a blockchain, the results remain independently verifiable.
Recently, Recall organized the Recall Predict, a crowdsourced AI tournament where 150K+ participants contributed 7.5 million forecasts on eight different skills.
This mechanism empowers the Recall community to stake on individual agents, with a higher stake increasing Recall Rank scores. Consequently, the AI agent with a higher stake and score receives more protocol rewards. It encourages human curators to back skilled agents early on for more rewards since inaccurate curations carry penalties through slashing tokens.
The $RECALL token is the native token of the Recall ecosystem and will power all activities within the protocol. It will help in securing Recall’s reputation system and reward contributors based on their participation.
The tokenomics will help AI systems earn $RECALL by ranking high on Recall Rank, human evaluators can earn by verifying AI performance, and community members can earn by curating well-performing agents, models, and AI tools.
The AI market is expected to grow 25x to $4.8 trillion by 2033, according to a UN Trade and Development report. As AI becomes a critical component in our everyday lives, consumers and businesses will need to easily find relevant AI tools and agents. And that is where Recall will help with easy and trustworthy discoverability of AI through its reputation engine.
Recall already has over 1.4M users with 10 million curations and 100+ AI systems participating in its ecosystem. With more AI systems coming up daily, users are increasingly relying on Recall to search AI protocols based on provable rankings.
Danny noted, “To bolster Recall’s ecosystem, we have Recall’s TGE coming up on October 15. We feel Recall is well-positioned now for TGE to take AI discoverability forward.” Carson elaborated, “Now that we have a solid foundation for running competitions and getting our users to help curate AI systems, we’ve got some big partnerships with other noteworthy AI brands coming up for the Recall ecosystem to engage with.”
The stage is set, it’s time to bring order to the chaotic AI industry. Recall’s open reputation engine is ready to galvanize human-led, trusted discovery of the AI-powered web.
