Deepgram transcription issues with names and common words

[ak-nextdim]

We are facing a lot of transcription issues with Deepgram in our voice agent setup. Most of the issues are related to incorrect or missed transcriptions, especially for names and short user responses.

• Names are often misheard and transcribed as entirely different names
• Words like “transfer” are being transcribed as “dancer,” “fantastic,” or “delta”
• Words like “maybe” are being transcribed as “navy”
• Short words like “sick” are sometimes not registered at all
• In some cases, entire patient names are not captured

We already provide key terms for some known names and terminology, but we are still seeing incorrect transcriptions.

Has anyone faced similar issues with Deepgram, especially in phone-call / voice-agent scenarios? Are there recommended settings or configurations that can improve recognition for names and words?

Our voice agents are built on LiveKit, and we’re using Deepgram as the STT provider.

[deepgram-community[bot]]

Thanks for asking your question. Please be sure to reply with as much detail as possible so the community can assist you efficiently.
_{Consider joining our Discord community for more opportunity to engage with your fellow Deepgram users. You can earn points which can be redeemed for cool stuff by being active in our communities!}

[deepgram-community[bot]]

Hey there! It looks like you haven't connected your GitHub account to your Deepgram account. You can do this at https://community.deepgram.com - being verified through this process will allow our team to help you in a much more streamlined fashion.

[deepgram-community[bot]]

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).
• A request ID that triggered your error or issue.

[ak-nextdim]

Python