We are evaluating whether we should bring our own provider keys to VAPI or if we should use the VAPI resources for our agents. We understand that VAPI probably has entreprise contracts that bring costs and latency down, but we can't find the information on how much of a difference it would be. Is there information anywhere on what these cost reductions would be?

Vapi passes through provider costs "at-cost" when you bring your own keys (BYOK), and only adds a small platform fee for orchestration. However, the documentation does not provide a public breakdown or comparison of cost or latency differences between BYOK and using Vapi's managed resources. For specific pricing details or to understand potential enterprise contract savings, it's best to contact Vapi's sales team directly for a custom quote or more information. [Learn more about Vapi pricing](https://docs.vapi.ai/pricing) Sources: - [Vapi FAQ At-cost and pricing](https://docs.vapi.ai/faq) - [Vapi Glossary At-cost](https://docs.vapi.ai/glossary#at-cost)

Hey Lucas, we don't have any difference in-terms of latency for enterprise customer and regular customer. Also, you can find the details for pricing over here ^^

The pricing details are not found on that page. VAPI says pricing is "at-cost" but the actual cost is not disclosured. For example, elevenlabs pricing is different across tiers. That means "at-cost" is a huge range of possible values. How much is the "at-cost" cost using VAPI vs BYOK?

You can find the cost details for each call in the call logs section. Just select any call_id to see the exact cost breakdown. Regarding the at-cost pricing, we are on an enterprise plan that provides us with a significant discount, and we are extending the same pricing to you.

Thanks for the info. We are using BYOK currently so we don't actually see the breakdown costs, and they're not listed in the docs. We'll kill our provider tokens and compare, but it may be worth adding a rough breakdown on the enterprise token discounts to the docs 😄

Hey Lucas, thanks for bringing this to our attention. It's a great idea that'll reduce questions about pricing and comparing key usage. Can you raise a feature request here: https://roadmap.vapi.ai