Skip to content

docs: add Marimba custom transcriber guide - #1426

Open
solanovisitor wants to merge 1 commit into
VapiAI:mainfrom
Charcot-AI:docs/marimba-custom-transcriber-v2
Open

solanovisitor wants to merge 1 commit into
VapiAI:mainfrom
Charcot-AI:docs/marimba-custom-transcriber-v2

Conversation

@solanovisitor

Copy link
Copy Markdown

Description

  • Adds fern/customization/custom-transcriber/marimba.mdx, a guide for running Marimba (Apache-2.0, open weights; paper: Marimba: A System One Model for Speech) in a custom transcriber bridge to decide when the caller's turn is over, alongside any speech-to-text model.
  • Adds the page to the Custom transcriber section in fern/docs.yml, after Gradium.
  • The bridge is the Vapi example in the Marimba repository. It runs against marimba-audio 0.1.2 from PyPI, reads only the customer channel (channel 0) of the documented stereo linear16 stream, and returns transcriber-response messages with partial and final transcripts.
  • The guide pairs the custom transcriber with a custom-endpointing-model route in startSpeakingPlan, using a matching per-call key.
  • Model numbers come from the model card: turn-end AUROC 0.930 on the Smart Turn v3.2 test set (Smart Turn v3.2 scores 0.982), and 30 languages.

Testing Steps

  • Run the app locally using fern docs dev or navigate to preview deployment
  • Ensure that the changed pages and code snippets work

Co-Authored-By: Solano Todeschini <solano@voahealth.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant