Pro tip to get your costs down FYI, I've made the tests last month of using my own OpenAI key instead of the vapi standard one for one small client. My LLM costs drastically reduced, as VAPI doesn't account for "cached input", although it's > 50% of my input tokens (see graph) and 4 times cheaper. https://cdn.discordapp.com/attachments/1211484481848873020/1412470214653644853/image.png?ex=68b868e5&is=68b71765&hm=c28af88329282af0e488fc4968b4c23afeb3ad85c710f7c1fb4cc5ccaaaa5a2f& https://cdn.discordapp.com/attachments/1211484481848873020/1412470215337578567/image.png?ex=68b868e5&is=68b71765&hm=05f0439936f7f454622505b640be63d27cf07f426edf58317117d5e3d86a8c7c&