I Compared a Bunch of AI Note Taking Apps for Therapists So You Don't Have To
And I learned a lot about privacy and security for digital mental health apps in the process
I’ll start by saying that this was a monster of a post to write, and I wanted to do it well because I think the information could be helpful for both therapists and curious clients. I began by trying to understand how these apps work, the privacy concerns associated with them, and how the apps manage and protect user data. I then generated a list of features that were important to me in a note-taking app, identified six apps specifically designed for therapists, and narrowed those down to two based on their security features. I compared the two apps against each of my criteria and ultimately selected a favorite. This post is not sponsored, and I did not receive any special deals to try out the software.
If you’re not interested in all the nitty-gritty details, you can skip to the end of each section for my bottom line in bold and then the final take at the very end.
Primer on the Purpose of Progress Notes
To get into the right mindset, I asked Claude what the purpose of therapy Progress Notes is, and I think it gave me a good summary:
Progress Notes in Psychotherapy serve several important purposes:
1. **Documentation of Treatment:** They record key details about each session, including the client's issues, goals, interventions used, and progress made. This helps in tracking the client's development over time.
2. **Continuity of Care:** Progress notes ensure that any mental health professional working with the client can understand what has been discussed and planned in previous sessions, maintaining consistency in treatment.
3. **Legal and Ethical Record:** They provide a legal record of the therapy process, which may be necessary for compliance, billing, or in case of legal reviews.
4. **Treatment Planning:** Notes help therapists assess the effectiveness of interventions and make informed decisions about future treatment strategies.
5. **Communication:** They can be shared with other healthcare providers involved in the client's care, facilitating a coordinated approach.
Overall, Progress Notes are essential for ensuring effective, ethical, and organized psychotherapy practice.
Thank you, Claude.
What I Want from a Note-Taking App
Okay, so with that in mind, I want a note-taking app that
Is designed for psychotherapists.
Proves that it takes privacy seriously.
Takes notes that accurately summarize and synthesize the session and need very little correction from me.
Is competitively priced.
Integrates easily with my EHR.
Is easy to navigate.
Is working on improving the technology by planning to offer better features in the future.
Takes psychotherapy notes that include important details that are too sensitive for progress notes but are details that I would like to remember in future sessions. This one could be a fantasy but I would like to see it. (For non-therapists, Psychotherapy Notes are different than Progress Notes in that they are not part of the official medical record and contain more personal details from the session.)
How I Chose Which Apps to Compare
I only selected AI note-taking apps explicitly designed for therapists and behavioral health specialists, rather than those intended for general medical care providers. Therapists' needs are specific enough to necessitate tailored software. I may revisit this someday and try out some bigger brands for general health care providers, but for now I’m sticking with the specialized software.
Once I decided to focus on apps specifically for therapists, I searched the internet and came up with a pretty long list:
Then I decided to define my bar for guaranteeing privacy and security. By law, any company that handles Personal Health Information (PHI) must be HIPAA compliant, but they all claim to be HIPAA compliant so that doesn’t narrow down my list at all. This isn’t the first rodeo for digital mental health, though, and it turns out that in the past, companies have claimed to be HIPAA compliant but were later found not to be. So I’d like a note-taking app that goes beyond making claims. Ideally, the app would be HITRUST certified, a gold standard certification that specifically tests HIPAA compliance, but for some reason, none of them are. SimplePractice is HITRUST certified, and I wonder if the note-taking app they are rolling out soon will also be covered by that certification. I’m not really sure why these other apps aren’t. I can only guess that the certification process is long and expensive, and these are scrappy startups that haven’t been able to allocate the resources to it. I hope they get certified soon, and I don’t think it’s crazy to wait to use an app until one of them does.
Not as good as HITRUST, but better than nothing is a SOC-2 certificate. A SOC-2 certificate focuses on data storage and processing security and demonstrates that the company has at least undergone some sort of third-party review. It is not specific to healthcare, though, and in theory, a company could be certified and still not be HIPAA compliant. But at least it’s something. Of the six apps above, only Blueprint and Upheal are SOC-2 certified (AutoNotes, Mentalync, and TheraPro say on their websites that they are working towards it). So Blueprint and Upheal it is.
So let’s see how Blueprint and Upheal compare on the criteria I outlined above!
Comparing Privacy and Security
I took a deep dive into privacy and security in digital mental health and learned a ton. I wanted to understand how the apps work, what the security risks are, and how these companies can protect client data. I also dug into what has gone wrong in digital mental health apps in the past. I came away with way too much information to share in this post, but I’m excited to focus on it all in a future post, as I believe this is an important concern for both therapists and clients. For now, though, here is a comparison of how Blueprint and Upheal handle security and privacy, specifically addressing some of the questions that arose during my research.
Are the companies transparent about how their apps are layered onto other apps, such as text-to-speech apps for transcription and large language models for language processing?
Neither company is particularly upfront about the fact that they work with third-party applications, but Upheal includes this information in the consent form that they provide for clients to sign, giving therapists permission to use the software during their sessions. Blueprint does not include this in their consent form, but it is there on their site if you dig into their security practices.
How do they handle data?
Both apps
Automatically delete and permanently remove audio recordings once transcriptions are produced.
Allow therapists to choose whether to retain or delete transcriptions, session summaries, and clinical notes at any time.
Allow therapists to set up “auto-delete” for transcripts and session summaries.
Offer a dictation mode, in which the therapist dictates a description of the session to be synthesized instead of recording it, making it easy to allow clients to opt out of being recorded.
The differences are that Blueprint allows you to enter sessions as “anonymous,” instead of linking personal health information to the recording and summary, and Upheal stores PHI in a pseudonymized format.
Is your data used to train their models?
Blueprint: Says “your data never trains our model.”
Upheal: Opting out from AI model training is only available in the paid version, not the free version. Upheal states that they never use sessions to train their platform without explicit permission from therapists and their clients. Upheal adds a voluntary second step of alerting clients to the use of their data for AI improvements.
How is their data encrypted? It should be AED-256 or stronger.
Blueprint uses AES 256 level encryption.
Upheaval uses “record level” encryption.
Bottom line: Blueprint is the winner here for me because they use AES-256 level encryption and never train their model with user data. However, I appreciate that Upheal explains their use of third-party apps in their provided consent form and that they store PHI in a pseudonymized format. But I just really don’t like how Upheal allows unlimited free sessions for therapists who allow their clients’ data to train Upheal’s model—more on that in the Pricing section.
Accuracy and Quality of Notes
To measure this, I first set up accounts with both Blueprint and Upheal. I was able to try both for free, although I understood that this would mean Upheal could use my therapy session to train their model. That feels a little strange, but I’m okay with it this time for this purpose.
I then recorded an audio file of my own personal therapy session (in which I was the client) with my therapist’s permission and uploaded it to Blueprint and Upheal, as well as Granola for comparison. Each app prompted me to select a note format from a long list of options, including a “custom-made” choice, and for each, I opted for the SOAP (Subjective, Objective, Assessment, Plan) format to keep it simple. Here is what I found:
Blueprint: Blueprint generated an accurate, succinct note with just the right amount of detail—enough to summarize all the important parts of the session and jog my memory when I reviewed the note later, but not so much that it felt excessive for a medical record. It effectively filtered out the chit-chat at the beginning and end of the session. Additionally, it provided a therapist’s summary, which I interpret as a sort of Psychotherapy Note, but it didn’t really go beyond what was in the Progress Note, as I had hoped. It also generated a client summary that a therapist could copy and paste for the client’s use, along with a separate section containing suggested assessments and worksheets that the therapist could assign to the client, as well as interventions that could be utilized. Lastly, it provided a transcript of the session. They say that if you don’t delete the transcripts, it allows the software to integrate information from past summaries into new sessions, although I wasn’t able to try this feature because I compared only one session. This is an interesting feature because I can see how it would improve the notes over time, but it also introduces additional privacy risks.
Upheal: Upheal generated a comparatively longer Progress Note than Blueprint. In addition to the SOAP sections, it also contained an MSE (Mental Status Exam) section, a detailed Interventions section, and a Risk Assessment section. The summary for the therapist is also quite a bit longer and breaks the session into topics, detailing the amount of time devoted to each topic. This section was organized in a way I did not find intuitive, and overall, I did not find it helpful. It also offered some analytics, detailing how much the client spoke versus the therapist (helpful), and how much the client’s sentiment was positive or negative (not helpful and kind of strange to me). It also had a transcript section where I could view the transcript for the entire session. I felt that the Upheal summary provided too much information in both sections, but especially in the summary for the therapist. Unlike in the Blueprint note, in the Upheal note, I did need to make one correction in the SOAP section.
Granola AI: I ran the audio through my favorite non-HIPAA compliant, non-healthcare app, just for comparison. I have to say that Granola’s summary was the best, demonstrating that there is still room for improvement in the specialized apps. It even seemed to recognize that I was in a therapy session and that the modality being used was IFS. It provided a summary of the entire session, with each section presented very succinctly. It offered just enough detail to capture the session accurately and jog my memory without overwhelming me with information.
Bottom line: In an ideal world, I’d have Granola’s accurate and perfectly succinct summary in a HIPAA-compliant app for therapists. However, in reality, when choosing between Blueprint and Upheal, Blueprint’s concise notes appealed to me more, along with its “Assist” features that provided suggestions. I bet these suggestions will become increasingly helpful as the software continues to improve. The Upheal notes almost seemed to me to be less mature than Blueprint’s, as if Upheal hasn’t iterated enough times to find a truly effective formula.
User Interface and Integration with Electronic Health Records (EHRs)
My experience with onboarding and navigating each site was positive. I preferred Blueprint’s design aesthetics slightly more, but I found the ease of navigation to be comparable on both sites. Using both was intuitive, and I never felt frustrated. In each instance, the user interface seemed to fade into the background, which was precisely what I wanted, allowing me to concentrate on the work.
Both apps state they offer a browser extension that integrates smoothly with the therapist’s existing EHR (Upheal does so in its paid plans, while Blueprint provides this in all plans). The Blueprint website specifically highlights an advanced integration feature for SimplePractice, whereas Upheal offers such a feature for Zoom, but only at their premium level. I utilize Zoom Healthcare for my virtual meetings and SimplePractice for my EHR, making it a draw for me. Copying and pasting is always an option in both apps, regardless of the plan.
I expected to have more to report here, but my experience with both apps in this category was fine, and both offer comparable EHR integration.
Pricing
Okay, so up to this point, I’ve liked Blueprint more. They never use your client data to train their models. I like some of their security features more, and they disclose a high level of encryption. They also created a nicely succinct and well-tailored note and offered suggestions that I could see being helpful. However, once we factor pricing into the equation, things get a little spicier.
Blueprint:
Free Plan: Blueprint offers a free plan that provides up to 5 sessions per month and includes all the bells and whistles. This is great for testing out the product to see if it suits your needs, or using it on all your sessions for free if you have a very small practice. And unlike Upheal, they state that they do not train their models on any user data.
Basic Plan: At the Basic Plan, Blueprint removes the bells and whistles, providing only notes, treatment plans, and discharge summaries. This plan starts at $29 per month for 30 sessions and increases from there, costing you about a dollar per session per month (45 sessions for $45 per month, etc.).
Plus Plan: The Plus Plan adds back those bells and whistles that you may have tried in the Free Plan at a cost of about 30% more than the Basic Plan.
Upheal:
Free Plan: Upheal also offers what appears to be a generous free plan, but, as I mentioned earlier, there is a catch. While they provide basic notes and even the use of their video platform for an unlimited number of individual sessions (not couples though), they are using these sessions to train their model. It’s essentially a trade. They let you use their software for free, and in return, you and your clients allow them to train their model on your sessions.
Starter Plan: At this level, Upheal is not using your sessions to train their models. Here they offer notes for couples work, offer many different formats for notes, and have options for using virtual video platforms other than their own, as well as in-person sessions. This includes unlimited sessions at a cost of $29 per month (or $19 per month if billed annually).
Plus Plan: This plan adds bells and whistles, including the ability to give the software custom instructions for capturing specific information you’re looking for, advanced analytics, and a session map detailing cadence, talking by each party, etc., treatment plans, and treatment plan integration with notes, and Zoom integration. This plan covers unlimited sessions for $99/month.
Bottom line: I still prefer Blueprint, but if you need it for much more than 30 sessions per month, it is more expensive than Upheal (the exact cost difference depends on the number of sessions you require). I like that Blueprint’s pricing is straightforward, and that it has a pretty reasonably priced Basic Plan and Plus Plan. I’m not sure that the bells and whistles you get with the Plus Plan are worth the extra 30%, but they might be, especially if they improve over time as the software continues to evolve. However, if you need the app for more than 30 sessions per month and cost is a significant concern for you, then I think Upheal’s Starter Plan is perfectly fine.
I also just want to circle back to Upheal’s unlimited free plan (if you skipped to the bottom line, it may be worth reading about this option above). This would probably be awfully tempting to many therapists, especially those just starting out or those who are operating on a tight budget. On the surface, it seems fine. As long as the therapist is clear with the client that the app will train on their sessions, it appears reasonable. But I worry about how this actually works in practice. The therapist would likely go over how this works with the client in the first session, along with their practice policies, confidentiality limitations, etc. In this session, the client is often nervous and has a lot to process before getting to the urgent mental health concern that got them into the office in the first place. Would the client really understand that because the therapist is using the software for free, the software would train on the client’s data? And what exactly does that mean? Is it possible that Upheal employees will be listening to client sessions to give feedback on the model’s performance? I’m not sure, but I imagine so. Many therapists will be drawn to this free package, but I am not one of them. It just doesn’t meet my standards for client privacy.
What’s Coming Down the Pipeline
Blueprint is advertising that it will soon be offering an AI-assisted EHR. Having an EHR built from the ground up based on AI sounds exciting. I’ll be eager to try this out and see how it compares with existing EHRs, such as SimplePractice, built before the AI era. Eventually, I’d like my note-taking app and EHR to be one, and with SimplePractice rolling out its note-taking app soon, it seems like these will be the two competitors.
Upheal appears to be releasing an update soon that will offer modality-specific notes across all its pricing levels, including the free level, allowing therapists to choose notes that reflect their specific practice model. For their Plus Plan, they will soon be offering progress notes that incorporate treatment plans, interventions based on modalities, and automatically generated case snapshots, providing the therapist with a quick summary of the client. I think the interventions based on modality is an exciting update, and I’d love to circle back and try it out once available.
Bottom Line: Blueprint’s plan to offer an EHR soon is a pretty big plus, especially if you’re like me and hope to have your EHR and note-taking app be one and the same. But I also have to say I like Upheal’s plan to offer modality-specific interventions. I think this could be an important direction for these apps to take (a topic I plan to write more on later), and so I’d love to see what Upheal does with it.
Final Take
I think you probably know by now which app I like the most. But to sum it all up in one place, I like Blueprint more because:
The way it handles privacy and security seems a little tighter.
I thought the notes were quite a bit better than Upheal’s.
The ease of navigation and attractive design.
The advanced integration feature with SimplePractice (although I did not test this out).
Its straightforward pricing.
Its plan to offer an EHR soon, so I’ll be able to do everything in one place.
With that said, I can see Upheal being a good option for some therapists because
The pricing ranges from on par to significantly cheaper than Blueprint’s, depending on the number of sessions per month (assuming you opt out of training on client data).
It offers an unlimited free option, although it comes with a significant caveat, an option I don’t love for the reasons I discuss in the pricing section above.
Nerd Note
I used NotebookLM heavily to write this post. I could not have gone into nearly the same level of detail in comparing the first six apps and then the two. To use it, I uploaded every page of each app’s website and then asked it to compare a bunch of different features. It summarized the information for me and provided me with links to its sources. I then checked all the sources, finding some hallucinations here and there, and then used the accurate information to write the post. If I had done all of this on my own, the post would have taken me many, many more hours to write.