boundary

Hey, most agent frameworks manage the tool-call loop natively -- does BAML support something equivalent (yet)?

we are re-writing a new data generator pipeline where agent tool calls are required. We are evaluating the options between Baml and pydanticAI

hey! we're going to be releasing something hopefully quite soon for this!

something that lets you own the full loop and makes testing a lot better

but still give you all the benefits of tooling with BAML that you've enjoyed so far

we're going to show off the earliest version of this tmrw during AI that works: https://luma.com/evals-revisited

cool! Look forward to the presentation tmrw!