We built Zero Runtime so teams could ship realtime AI without spending months on the plumbing. This blog is where we'll share what we learn along the way — the architecture decisions, the things that broke, and the patterns we wish we'd known about earlier.
What we'll write about
- Engineering deep-dives on the systems behind realtime agents.
- Operational notes from running them at scale: observability, reliability, and rollout strategy.
- Product updates when something ships that changes how you build.
Why a static, file-based blog
Every post lives as an .mdx file in this repository. There's no CMS to
log into, no preview environment to babysit, and no separate deploy
pipeline. Drafting a post is a normal pull request — review happens in
the same place as code review.
If we wouldn't tolerate a CMS for our application code, we shouldn't tolerate one for the writing about that code.
See it in action
A short walkthrough of what realtime agents look like end-to-end:
And a sample of what one of those agents sounds like in conversation:
What's next
Real posts start landing here over the coming weeks. If you'd like a specific topic covered, let us know — half of what we write is in response to questions teams send us directly.
