Filler words not appearing/marked in transcript despite filler_words=true parameter

[fierce bear]

Thanks for asking your question. Please be sure to reply with as much detail as possible so the community can assist you efficiently.
-# If you haven't done so, ensure your Discord and Github profiles are linked to Deepgram so you can earn points to redeem on cool stuff just by being active!

It looks like we're missing some important information to help debug your issue. Would you mind providing us with the following details in a reply?

• The programming language you are working in (e.g. JavaScript, Python).

[runic copper]

I used the playground to test this for now.

[blissful spear]

Filler words should show up directly in the transcript when using filler_words=true. That said, we are aware of model shyness with filler words, and it's something that we'll be working on improving over the coming months. Nova-2 is slightly better than Nova-3 at filler words, and Flux is much better than either of them at transcribing filler words. With Flux there is no enabling or disabling, so filler words will always be transcribed.

[visual osprey]

Hey, are there any plans to allow disabling the filler words in deepgram Flux model? It would be a super helpful have them removed from transcription

[fierce bear]

⚠️ An error occurred while syndicating this conversation to our internal channels. @warm arch has been notified.
-# 1468582551000977535

[blissful spear]

we aren't as of yet planning to support stripping out filler words on Flux, though I'd love to hear about why that is important to you and your use case to inform future prioritization!

[visual osprey]

Hi! thanks for the response!

Our main use case is formal, structured live conversations such as screening interviews used to assess candidates. In this context, filler words like “um”, “uh”, “oh”, etc. tend to add noise rather than value.

From both sides:

• Candidates often feel uncomfortable seeing filler words reflected back in transcripts, as it can make them appear less articulate than they actually are.
• Screeners/reviewers typically focus on clarity, content, and reasoning, and filler words don’t contribute meaningfully to the assessment.

Being able to optionally strip or suppress filler words would help:

• Produce cleaner, more professional transcripts
• Improve candidate experience and confidence
• Reduce cognitive load for reviewers when evaluating responses

Happy to share more detail about the workflow or examples if helpful - and appreciate you considering it for future prioritization.