We’re actively exploring ways to add speech-to-text support to OpenEMR to help doctors document more efficiently and reduce manual note-taking.
If someone wanted to use voice input in OpenEMR today:
• What options are available?
• Has anyone worked on something similar?
• Are there any tools, experiences, or lessons learned, even unsuccessful
ones, that you’d be willing to share?
Any insights would be really helpful as we shape our next steps.
@luisuriarte
Hi Chris,
Thanks for your response!
Would love to hear more about your client’s needs and what you’ve seen work (or not work) so far.
Any insight would be super helpful as we explore this.
Thanks so much for jumping in. Really interesting to hear that you implemented this a few years ago!
Would you be open to sharing a bit more about:
What the technical integration looked like?
Any limitations or lessons learned along the way?
And whether your clients are still using it today?
We’re gathering as much insight as we can from past efforts to help shape what could come next.
Whoops — tagging fail on my part!
Thanks @luisuriarte for actually being the one who brought Chris’s post back into the spotlight. That thread really helped shape this whole discussion.
And @faster-than-human now that you’ve been double-summoned — would love to hear any thoughts or experiences you might want to throw in!
Just wanted to follow up and share that I’ve been reviewing earlier discussions around voice input and dictation in OpenEMR.
Over the years, there have been several interesting attempts, from experiments with Dragon and CKEditor plugins to open-source tools like Simon and DeepSpeech.
Here are a few past threads, listed chronologically:
• Dragon dictation setup (2009)
• Speech dictation experiences (2017)
• Open-source speech recognition (2017)
• Voice dictation software (2019)
• Voice to Text with AI revision assistance (2025)
Each of these threads shows different angles and motivations, but it’s not clear if any of them led to a long-term solution.
I’m curious, does the community still feel there’s a strong need for voice input in OpenEMR today?
Would love to hear your thoughts.
I have posted this video before. here is it is again. I have this implemented in version 7.0.2 (1). The way it is built is similar to what @sjpadgett did with nation note. It will attach to any text area only that is on a page. The revise button transmits the text to the AI with a prompt of what type of persona, voice, and tone to use for revising the text and populates it back into the text areas.
Thanks a lot for re-sharing this Sherwin. Really appreciate it.
Do you recall whether this approach was ever used in production, or if it stayed more of a proof of concept?
I’m trying to map out what’s been explored so far and understand what has actually worked or stuck.
Also curious, from your perspective, do you think the OpenEMR community is still interested in capabilities like this, or perhaps even broader AI features?
Jazmine, I won’t divulge to much. You can ask more questions. What tools did I use? I used a voice to text process that has been around for a long time. It is nothing new. I am not sure how to answer the workflow question.
The tech is attached to the LBF form. So, any LBF form that is generated with a text area will have this tech applied to it like the nations notes.
It was interesting that when I deployed the tech for a client to test and use. They don’t use the revise with AI as much as I thought they would. I don’t know how much they are using the speech to tech because it is enclosed within the application. It is not a third party service like the AI is.