Snips AIR Token Sale
Powering the Snips Private-by-Design,
Open Source, Decentralized Voice Assistant
Snips runs 100% on the edge, without any data being sent to the cloud. This guarantees Privacy-by-Design and protection against mass surveillance.
Everything from the wake word to the ASR, NLU, and dialog runs locally on the device to which users speak. Snips can run on any hardware platform and most operating systems.
The Snips assistant resolves queries on a Raspberry Pi in less than 200 ms, using around 100 MB of memory, and with an accuracy as good or better than cloud-based assistants.
Everything in Snips is modular and customizable: from the wake words to the ASR, NLU, dialog or TTS. You can run them together or individually on different devices.
Our NLU engine is already open source, and we are committed to open sourcing the rest of the platform in the near future.
Add a voice to your home while protecting your family's Privacy.
A Network of AIR Devices
AIR is a network of devices collaborating to form a mesh of AI assistants. This enables your entire home to be seamlessly voice-enabled while protecting your family’s privacy. Since Snips AIR runs locally without any voice data being sent to the cloud, no one can spy on you or your kids, keeping your home private as it should be.Discover Snips AIR
Decentralized Skill Store
The Snips Skill Store uses Snips AIR tokens to ensure anonymous, decentralized transactions between users and skill makers. Quality, ranking, reviews, and engagement are incentivized using a combination of staking and transaction fee sharing.Try Existing Skills
Anyone can easily build skills for Snips and earn tokens. Although the Snips AIR token is not yet available, almost 13,000 developers are already building free skills for Snips, making it the largest community of voice developers outside of Amazon and Google.
Through the online skill builder, developers can improve the quality of their skills by automatically generating high quality NLU training data using our human-in-the-loop algorithm. Developers pay community “workers” using tokens, with community “clerks” validating results. Tasks are prioritized according to the number of tokens the developer holds. Workers and clerks need to stake tokens to participate, and are penalized for providing a wrong result.
A decentralized, encrypted analytics engine for skills. Encrypted usage data is processed by clerks and aggregated cryptographically by the skill developer, who then pays the clerks via the blockchain. Clerks need to stake tokens to participate in the network, and will be penalized for producing an incorrect result. The engine is already open source.
Decentralized machine learning to train AI models with real user data, while protecting user privacy. Users get paid for participating in a "federated learning" network, with clerks getting paid to process the partial models. Clerks need to stake tokens to participate in the network, and are penalized for producing an incorrect result. Users get paid without staking, but proportionally to how many tokens they hold overall.
Learn More About Snips
Snips is an established company at an advanced development stage, having shipped products with dozens of features, having an active community and a growing enterprise business. The management team has been collaborating successfully for several years.