Rime MCP

[x] cli demonstrating that buffering actually reducing time to first voice play [x] add env vars to configure (guidance, who_to_address, when_to_speak, voice) [x] clean up extra stuff from readme [x] add examples of how it might be set up on readme [] try it out in cursor [] demo video Notes: - no response format? https://docs.rime.ai/api-reference/endpoint/json-mp3 - voices doesn't actually have a list of voices? https://docs.rime.ai/api-reference/voices -- need to hit a link to load them, maybe limitation of mintlify - my account rolls back to "no plan" and then i select free and it works again