Story Protocol Digest #074
May 5 - 12, 2025
Numo went from 110,000 to 210,000 contributions in a single week, growing roughly 40% day over day. Every recording rights-cleared and licensed onchain at the moment of contribution. Vietnamese launched, Bengali hit its target, and the platform is expanding beyond voice into hard-to-source data that can’t be scraped. Alongside that, a new CDR agent skill makes encrypted data workflows accessible to any compatible AI agent in a few lines of code.
The data supply problem for AI is getting a concrete answer. Week by week.
Numo Surpasses 210,000 Voice Recordings Across Early Language Rollout
Numo has surpassed 210,000 rare voice recordings from more than 15,000 contributors, with growth accelerating despite currently supporting only four languages. Every recording submitted through the platform is rights-cleared and licensed onchain at the moment of contribution, reinforcing Story’s approach to compliant AI data infrastructure. The rapid growth highlights increasing demand for real-world multilingual datasets that traditional web scraping methods cannot provide. Additional languages and new task categories are expected to launch soon as the platform expands beyond its initial rollout.
CDR Skill Makes Confidential Data Workflows Accessible to AI Agents
Jacob Tucker released a new agent skill and end-to-end examples for working with Confidential Data Rails. The skill allows developers to build AI agents capable of sharing and licensing private data onchain without exposing the underlying information. Developers can install the skill directly or provide the GitHub repository URL to compatible agents for immediate integration. The release lowers the barrier to experimenting with encrypted data workflows and demonstrates how CDR can operate within agent-based systems.
Vietnamese Language Tasks Launch on Numo
Numo has expanded into Vietnamese voice data collection, marking another step toward building multilingual, real-world AI datasets. Vietnam’s rapidly growing digital economy and mobile-first internet culture make it a critical environment for future AI systems, yet high-quality Vietnamese voice data remains limited. The language presents unique technical challenges because of its tonal structure, strong regional dialect variation, and pronunciation differences across Northern, Central, and Southern speech systems. Existing AI models are heavily biased toward Northern Vietnamese, leaving significant gaps in real-world understanding. By sourcing conversational and consented data directly from contributors, Numo aims to improve how AI systems interpret natural speech patterns in Southeast Asian environments.
Numo Growth Accelerates as Long-Tail Data Collection Expands
Story President and CPO Andrea Muttoni shared early traction metrics from Numo’s first week following launch with Poseidon. The platform recorded more than 210,000 contributions from over 18,000 contributors, with submissions growing roughly 40 percent day over day. Each contribution is registered as IP and licensed onchain through Story infrastructure at the moment it is created. Vietnamese was prioritized because of its tonal complexity and regional variation as important stress tests for voice AI systems. Beyond voice, the platform plans to introduce additional task categories focused on hard-to-source, real-world data that cannot easily be scraped from the internet.
ip.world Introduces Treasury System for Community Tokens
ip.world has launched a treasury system designed to give token communities direct control over trading-generated revenue. Unlike traditional memecoin ecosystems where fees are captured primarily by platforms or early participants, each token on ip.world now includes a treasury funded automatically through trading activity. Holders can submit and vote on proposals to deploy treasury funds, including wallet transfers, token burns, or asset swaps. The feature introduces governance-driven community economies where participation and ownership are more closely aligned.
Confidential Data Rails Continues Expanding Developer Access
Confidential Data Rails remains fully available on Story testnet, with live applications for private storage, encrypted data exchange, and developer experimentation. The release reinforces Story’s broader push toward making sensitive data economically usable while preserving confidentiality and programmable control. Developers can explore current implementations and begin building through the public testnet environment at
Poseidon Shifts From Scale to Precision in AI Data Collection
Last year, Poseidon demonstrated that distributed AI data collection could scale. This year, the focus has shifted toward precision and quality targeting through Numo. The team pointed out that the useful shelf life of AI training data is often shorter than expected, meaning value increasingly depends on sourcing the right data rather than simply collecting more of it. Current efforts prioritize targeted markets, high-signal datasets, deliberate intake caps, and stronger quality standards. Bengali voice collection has already reached its target, while Hindi, Tamil, Telugu, and Vietnamese datasets remain open for contribution.
Story Community Welcomes New Seekers and Adepts
The Story community continues to expand through new contributor recognition across both Seeker and Adept tiers.
@mrthanh1978 joins as a Seeker following support for projects such as MusicByVirtuals, ip.world, and Poseidon through artwork and content contributions.
@QuyTu140498 has maintained a strong presence within the Vietnamese community through educational content and ecosystem engagement.
@melophile0646 advances to Adept after months of producing highly detailed graphs, infographics, and explanatory content around Story.
@Khan_uk17 has become a recognizable voice across community discussions and activities, while
@ranas1814 moves into the Adept tier following consistent high-effort contributions and active participation in OGs Lab.
This newsletter is brought to you by Chainflow — an independent, self-funded validator supporting decentralized PoS networks.