Our approach to sourcing data attributes is based on the token type and attribute category. We categorize our data collection into three main techniques: API only, API + Manual, and Manual. The primary sources for token content are the token website, white papers, and for L1 or L2 networks, the principal block explorer associated with the blockchain. For economic data (e.g., total supply, market cap, price), we source from multiple reputable data aggregators. We use time-specific data (previous day) for consistency in disclosure-driven data, with processes to compute and normalize output on a per-attribute basis. For network and protocol activity, we utilize on-chain data from node aggregators, as well as indexed data from major blockchain data providers. In cases where on-chain data is not immediately available, we rely on established analytics platforms. For stablecoins, we extract data from periodic reports detailing underlying asset allocations, as these are primary sources from the issuers. Announcements and events are sourced from crypto news sites, mainstream media, and social media platforms. We leverage AI tools to summarize documents and text, always providing the original sources alongside the summarized output. Data aggregation is a combination of manual inputs and automation, depending on the source and attribute type. The reliability of sources is determined by their methodologies, update frequencies, and reputation among clients.

Our verification and validation process is systematic and thorough, addressing the diversity of our data sources. For API-sourced data, we conduct regular validations (monthly or bi-monthly) including cross-checking against other reliable sources, verifying accuracy against blockchain records, assessing consistency with market trends, and checking for logical errors or inconsistencies. For manually collected data, we employ a multi-faceted approach including expert reviews, cross-validation using secondary sources, sentiment analysis, and timeliness checks. All textual data undergoes initial quality checks for completeness, adherence to formatting guidelines, and identification of obvious errors. Content validators review the data, comparing it against multiple sources to ensure reliability. We have defined thresholds for each data element, and any variations beyond these thresholds trigger a manual verification process. The actual process of normalization or averaging is published for transparency and to allow independent verification. We use APIs for automated integration but also provide links to data aggregator sites for direct verification where possible. When using AI tools for summarization or synthesis, the output is manually verified against the original sources. For single-source data, we clearly indicate the source and are developing a curated list of dependable sources through a governance process.

Our commitment to accuracy, reliability, and timeliness is reflected in our comprehensive methodologies. We continuously refine our approach, incorporating user feedback and industry best practices, to adapt to the dynamic nature of the digital asset landscape. This policy and our methodologies are regularly updated and version-controlled for transparency.

Note: The sources, approach, and methodology defined here are currently being followed for our pilot program. We will continue to refine these as we advance our disclosure service.