> For the complete documentation index, see [llms.txt](https://whitepaper.silencio.network/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://whitepaper.silencio.network/silencio-introduction/how-we-source-data.md). # How We Source Data One network, one engine, every dataset. Silencio turns everyday devices into a global field-data engine. There is no proprietary hardware and no recruiting agency in the loop: contributors use the smartphones and browsers they already own, and the network reaches anywhere people are. **The same engine that mapped the world's noise now captures its voices**, and it can be retargeted to whatever dataset AI needs next. ### Who contributes, and what they produce **Millions of contributors across nearly every country on earth** produce the data buyers cannot find anywhere else, in **thousands of languages and dialects**, in real conditions. Through **Silencio Voice AI**, contributors deliver three core data types: **scripted speech**, read from prompts to create labeled, structured corpora across languages and accents; **natural conversation**, unscripted dialogue for training conversational and dialogue models; and **command data**, short task-based utterances for assistants and robotics control. Around the voice core, the network continues to capture **ambient and environmental signal** through Silencio Sound Check, the noise intelligence layer where the network began, and is extending into **egocentric, first-person data collection** for embodied AI. ### How quality is assured Scale is worthless without quality, so the network polices itself. Contributions pass through **layered validation: automated screening for technical quality, human peer review by other contributors, and expert checks on sampled data**, before anything reaches a buyer. Contributors build reputations over time, and rewards follow accuracy, which aligns every participant with the quality of the catalogue. The result is field data with studio-grade reliability, delivered as **model-ready datasets with transcription and validation included.** ### How consent and payment work Every contribution is opt-in and anonymized, and every consent is recorded as a **verifiable, tamper-proof entry tied to the contribution itself**, creating the audit trail the EU AI Act demands and scraped or synthetic data can never have. Contributors are paid in **USDC base pay** for accepted work, real, stable, withdrawable value, with **$SLC rewards, raffles, and bounties** as the upside layer for consistent participation. People are not data subjects here. **They are the supply side of the business, and they are paid like it.** ### What buyers receive Buyers access the network three ways: an **off-the-shelf catalogue of hundreds of thousands of hours and growing**, **on-demand collection** in any language, demographic, or acoustic condition, and **transcription and validation** as a service on top. Delivery is enterprise-grade and provenance is provable on demand. **Frontier AI labs and Fortune 100 companies already train their voice AI on data sourced from this network.** ### The flywheel The economics close the loop: contributors supply data, buyers pay for datasets, and revenue flows back, to the contributors who earn from it and to the token economy that tightens around it. **More contributors mean better coverage, better coverage wins bigger buyers, bigger buyers fund more contributors.** The network does not just scale. It compounds. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://whitepaper.silencio.network/silencio-introduction/how-we-source-data.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.