Skip to content

[Suggestion] Flux in-band reconfiguration: update keyterms and language mid-stream without reconnection (Python) #604

Description

@deepgram-robot

What to build

A Python recipe demonstrating Deepgram Flux's unique in-band configuration capability — dynamically updating keyterms, language hints, and end-of-turn threshold during an active streaming session without closing and reopening the WebSocket connection.

Why this matters

Deepgram Flux is the only STT provider that supports updating transcription configuration mid-stream. This is critical for voice agents that need to adapt to conversation context — adding domain-specific keyterms after detecting a topic, switching language detection modes, or adjusting endpointing sensitivity based on user behavior. Despite being a unique competitive advantage, no published recipe demonstrates this capability.

Suggested scope

  • Python, using deepgram-sdk
  • Deepgram Flux streaming STT
  • Demonstrate mid-stream updates: keyterms array, language_hints, eot_threshold
  • Show before/after transcription accuracy improvement when keyterms are added
  • Include timing measurements showing zero-downtime reconfiguration vs reconnection approach

Acceptance criteria

  • Runnable with minimal setup (clone, add API key, run)
  • README explains the pattern clearly
  • Uses current SDK version
  • Demonstrates at least 3 configuration parameters updated mid-stream
  • Shows measurable accuracy improvement from dynamic keyterm injection

Raised by the DX intelligence system.


Queued by PM — Engineer will pick this up as a priority:user recipe.

Metadata

Metadata

Assignees

No one assigned

    Labels

    action:generateQueue action: generate code examplespriority:userUser-submitted request — prioritizetype:queueIssue is in the automation queue

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions