You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Question: Is this something we can accomplish in a generic way? Can we write one thing, and then possibly add a configuration parameter to turn it on for each model - or maintain a database of models / providers that have that capability?
Or is this something we need to implement on a provider by provider basis? If that's the case instead of asking for "all LLM" we might want to start compiling a list of precisely what is wanted. @prd-tuong-nguyen@xiaoshyang@VikashLoomba could you elaborate a bit on what you are trying to do with tools and what providers and APIs and models you would like tool support on?
@shermanhuman I think we should prioritize support for the OpenAI Compatible format, as nearly all LLM serving frameworks and providers support this format. In other words, if we support OpenAI Compatible, we support nearly "all LLM"
I agree with what @prd-tuong-nguyen said. If we can enable OpenAI based tool calling (perhaps by a per-model setting "tools_enabled" in config.json) first, that would be most useful.
Validations
Problem
We can currently only use the tools with Anthropic
Solution
Permit all LLM to use MCP's tools.
The text was updated successfully, but these errors were encountered: