From 60 seconds to under 4: a two-day WAL migration
How we took POST /v1/events from 2.11 rps and a 60s p99 to 108 rps and a 3.6s p99 in 48 hours, uncovered eight production bugs, and cut the global flip live.
Read post →
Blog
Engineering deep dives, product decisions, and practical guides for teams building event-driven customer communications.
Jinja2, Liquid, and Handlebars were designed for developers. We needed a template language that marketing teams could read, write, and edit without filing engineering tickets. Here is how NLT works and why we made the trade-offs we did.
Coming soon
Sandbox environments hide real problems. Broken rendering, incorrect personalization, and deliverability issues only surface in production — unless your test mode uses real delivery. We built test mode differently.
Coming soon
From event ingestion to template rendering to delivery — a technical deep dive into the Synapse pipeline. Pub/Sub, async workers, dead letter queues, and the architecture decisions behind sub-200ms event processing.
Coming soon
If your templates are written in Jinja2, Liquid, or Handlebars, here is how to migrate them to NLT — with before-and-after examples for variables, conditionals, loops, filters, and currency formatting.
Coming soon
Sign up for Synapse and get product updates as we publish them.