Tokoro – open, signed event protocol with LLM crawler (github.com)

Tokoro has launched an innovative open-source, self-hosted event protocol that utilizes large language models (LLMs) to streamline the discovery of local events. The platform addresses the fragmentation of event information across various websites and social media by allowing curators to extract event data seamlessly from any web page. Using simple tools such as a browser extension or CLI crawler, users can indicate event listings and simultaneously extract and publish them into a shared geographical calendar. This process significantly reduces the time and effort required to compile event data, enabling communities—like jazz fans nationwide—to collaborate efficiently. This development is significant for the AI/ML community as it exemplifies the potential of LLMs beyond conventional applications, illustrating their capability to enhance data extraction and event aggregation tasks. Tokoro’s architecture leverages Cloudflare Workers and D1 (SQLite) to provide a responsive API, ensuring that events are signed for authenticity without the need for user accounts or passwords, as identities are verified through Ed25519 keypairs. By enabling users to query events based on location, time, and category, Tokoro not only fosters community engagement but also opens doors for developers to create applications around this curated data, further driving innovation in event discovery and social interaction.