Changelog #0009

Happy Friday!

Nothing user-facing to report in this week’s changelog, so hang tight.

We continue to invest in internal tooling: evals, infra rewrite, and prepping last week’s automatic template induction POC for production.

For fun, here’s a shout-out for some of the tools we’ve been using and loving:

• Bun for runtime and tests.

• E2B for TS sandboxing.

• Braintrust for evals.

• Vercel AI SDK for structured generation.

• And of course, Butter for caching tests and evals - it’s really sped up our iteration cycles to have LLM requests return 20x faster.

Stay tuned next week for more updates!