We use Factory because they've made it very easy to have long-running agents, long conversations with good context managements, and switch between models (e.g., Claude for building, GPT for reviewing).
Exa has been super easy to integrate as a back end search provider for our chat. They've also been very helpful in working with us to meet our security requirements.
Cubic was really good at finding subtle issues in our PRs that human and other AI reviewers could have easily missed. Because we have a lot of security critical code, having a reliable code reviewer with a good security sense like Cubic and a low false positive rate has been super helpful!
Tinfoil
