Articles in the Engineering category.
How Convor treats support attachments: size limits, MIME allowlists, storage permissions, malware scanning hooks, message links, and retention.
How Convor uses visitor identity, page context, UTM data, consent, browsing history limits, and dashboard UI without over-collecting.
A look at the embeddable loader: singleton initialization, iframe isolation, postMessage checks, CSS validation, route tracking, and cleanup.
What privacy-first means in product code: less data by default, organization boundaries, visitor identity limits, export and deletion flows, retention, and audit logs.
A practical walk through the Convor message path: Fastify writes, PostgreSQL as the source of truth, Centrifugo channels, unread counters, typing signals, and reconnect behavior.
