Congrats on the launch, I'm a happy user of Dagger. One use case I've found for it besides CI was building an eval sandbox for testing different coding agents' performance using CLIs, APIs, and coding tasks.
It helps a lot with layering different permutations of dependencies (i.e. test Claude Code vs Codex vs Pi, or forward port access to the docs or no docs access) and you also benefit from the layer caching so evals take a lot less time.
One piece of constructive criticism I have for Dagger: while the API and CLI are beautifully made, the documentation and the webapp scare the normies and could be improved. I've also had some trouble with outdated documentation in the past but I'm sure 1.0 is meant to be a turning point in that with the new docs rework and all.
Was gonna post this, you beat me to it!
Congrats on the launch, I'm a happy user of Dagger. One use case I've found for it besides CI was building an eval sandbox for testing different coding agents' performance using CLIs, APIs, and coding tasks.
It helps a lot with layering different permutations of dependencies (i.e. test Claude Code vs Codex vs Pi, or forward port access to the docs or no docs access) and you also benefit from the layer caching so evals take a lot less time.
One piece of constructive criticism I have for Dagger: while the API and CLI are beautifully made, the documentation and the webapp scare the normies and could be improved. I've also had some trouble with outdated documentation in the past but I'm sure 1.0 is meant to be a turning point in that with the new docs rework and all.