MCP & integrating AI chat & managing token costs

Ask us anything How do I?
1

I watched the webinar with WeWeb and Xano on how to implement a MCP and was quite blown away with how easy it seems to setup and the potential for offering my users a fast self-service way of analyzing their data. I had asked during the demo about how to leverage the use of multiple LLM’s depending on context/task, and the reply was that I should look into services that package subscriptions to the top models.

Anyone have experience implementing this? Seems like all of these ai packaging services are geared toward individual users. Any specifically for integrating into apps?

Also don’t want runaway costs for AI. Wondering if a better solution is offer the hooks for my users to supply their own API keys for AI model, which would likely mean I would need to narrow to 1 or 2 models?

2

Hi Eric, this is a super interesting question which I don’t quite have the answer.

Let’s hope someone from the community can share their experience with us on that issue.

3

https://fireworks.ai is your friend. Blazing fast inference for 100+ models.

1 Like