SOAPsuds
SOAPsuds, from NoteIQ LLC in Ashburn, Virginia, is an ambient medical scribe that turns a visit into a SOAP, DAP or progress note. Clinicians record in person, capture both sides of a telehealth call by sharing tab audio, upload an existing recording or dictate when recording does not suit the visit. A visit recorded without a connection stays encrypted on the device until it uploads, so a clinician can record on a phone and finish on a computer. Custom templates come from an example note the clinician uploads, and the Custom plan adds ICD-10 and CPT code suggestions, differential diagnosis suggestions, a documentation improvement module and white labeling.
The clinician reviews, edits and signs each draft, then copies the whole note or one section into the EHR or exports a PDF or Word file, since there is no direct EHR connection. Confidential adolescent content stays out of whole note exports, and an activity log shows who opened each visit. Audio, transcripts and notes are deleted 30 days after the visit.
Deepgram, OpenAI, Google Cloud and MongoDB are named processing partners, and a BAA is accepted in the app before the first visit. A 20 note trial needs no card. The pricing page lists Individual at $149 a month for unlimited notes, while the help center lists $40 for 40 notes and Unlimited at $100.
Capability Axes
An AI Health Index grade measures what a buyer can verify from public sources on the date shown. It is not a rating of how good the product is. A vendor can build an excellent system and grade low on an axis because it publishes nothing an outsider can check. How grades read
Speech recognition and note generation are the product. A clinician records a visit, uploads an existing recording or dictates, and the system turns that audio into a structured SOAP, DAP or progress note in the clinician's chosen template. Code suggestions, the documentation improvement module and differential diagnosis suggestions all work from the same generated content.
Templates, practice roles, the activity log and the copy and export tools support the workflow, but without transcription and language modeling nothing of the product remains. Its value rests on pulling clinically relevant details out of a conversation and arranging them into a draft the clinician can edit and sign, which makes it an AI product rather than a charting tool with an assistant attached.
Every note stays a draft until the clinician reviews, edits and signs it. Each section can be changed before it is copied into the EHR, edits save as the clinician works, and code suggestions wait for review. The help center describes the service as an assistant that writes a draft and does not make clinical decisions.
Marketing pairs a claim of no assumptions or inaccuracies with a 98% alignment figure, and neither removes the review step. No confidence threshold says when the system should stop short of a note or ask for more information, so the clinician supplies the final judgment while the rule for withholding incomplete content goes unstated for both notes and differential suggestions.
Deepgram, a speech recognition company, and OpenAI appear among the processing partners alongside Google Cloud and MongoDB. Natural language processing reads the conversation, and context aware models arrange it into note formats. Templates are built from an example note the clinician uploads, checks section by section and approves, and the product adapts to clinician corrections.
Model versions, how corrections change later behavior and the training data behind the medical terminology claim are not given, and no update schedule exists. The named suppliers say more than a generic AI claim, while the exact transcription engine and generative model configuration used for a given visit stay unclear.
Patient information passes through Deepgram, OpenAI, Google Cloud and MongoDB for transcription, generation, storage and data management, so the named list covers model providers as well as the cloud host. Product analytics track a few onboarding steps, such as creating an account and starting a first visit, and never receive patient details or note text.
The help center's answer on AI training says the company cannot yet confirm in writing the terms for every service it uses and offers to email current terms on request. No product by product subprocessor register ties each data type to its recipient and governing terms, so how long each party keeps audio or text, whether a model can be swapped and which safeguards follow a transfer remain open.
Marketing claims 98% alignment with spoken details, a 43 second average chart closure, more than two hours saved a day and a 30% revenue increase. No denominators, observation periods, reference standard or comparison group sit behind those figures. A family medicine testimonial names Dr. Emily Chen without an institution or study design, and the homepage quote about getting evenings and weekends back is unnamed.
The revenue calculator produces estimates from a user's own inputs rather than measured customer results, and app store install counts describe adoption rather than benefit. None of this tests the current models for accuracy, omission of clinically important details or the quality of differential suggestions.
Each visit's audio, transcript and note is deleted 30 days after the visit, and a clinician can delete a visit sooner. BAA acceptance records and the immutable activity log are kept for at least six years. Product analytics exclude patient details and visit text, and a confidential adolescent section is left out of whole note copies and exports to limit disclosure through patient portals.
The privacy policy permits use of data for AI model improvement and training, while the help center says training terms are not yet confirmed for every service. Removal of identifiers before model use, deletion from downstream model services and the steps taken after a harmful generated note are not described.
A solo clinician accepts the BAA at registration or after sign in, and a practice administrator accepts it for the whole practice before the first visit. Accounts configured by the vendor use the agreement in their contract. Settings keep the accepted version, signer and timestamp with a printable record, retained for at least six years.
The help center describes encrypted storage, 30 day deletion and an activity log showing who opened each visit, and leaves the judgment of whether a practice's own use meets HIPAA with that practice. The full agreement is available in the account or on request, so subcontractor duties and module scope cannot be read before sign up.
Visit records are stored encrypted, offline recordings stay encrypted on the device until upload, accounts sign out after an hour without activity, roles restrict practice access and an immutable log records who opens each visit. Sign in codes and password resets go to the clinician's email, and the help center asks users to protect that inbox.
The privacy policy mentions a HIPAA compliance audit and a SOC1 policy without naming an auditor, period, scope or report, and no SOC 2 or ISO certification is held. Startup program ties with Google and Microsoft do not certify the company's own operations. The controls are specific, and independent assurance of them is what is missing.
The service terms limit SOAPsuds to transcription and organization for licensed professionals and exclude diagnosis, medical advice and treatment, and the help center says it does not diagnose, refer, set urgency or make clinical decisions. The homepage and the Custom plan, however, sell differential diagnosis suggestions drawn from the encounter.
That puts a decision support feature inside a product whose terms describe documentation only. No device clearance or specific regulatory exclusion accompanies the diagnostic suggestions, and clinician review limits how they are used without settling which side of the line between drafting and decision support the feature sits on.
Clinician review and signature, editable sections and a documentation check provide oversight of each generated record. Privacy controls keep confidential adolescent content out of ordinary exports and log who opens or changes a visit, and practice administrators assign roles for clinicians and office staff.
Ethical AI messaging commits to fairness and transparency, but no model evaluation compares performance by accent, language, specialty, age or encounter complexity, and the 98% alignment claim has no subgroup breakdown. The interface and help center run in English only. Governance therefore rests on each practice's review process and stated principles rather than a tested framework for model behavior.
Clinicians can correct every section before signing and exporting, and the activity history records changes and access. Visits and the whole account can be deleted, with account deletion confirmed by an emailed code. Support can investigate a problem from the visit date and time alone, without patient details.
The service terms disclaim completeness and accuracy and exclude responsibility for decisions based only on generated notes. No standard duty promises compensation or correction of a harmful clinical conclusion, so a practice's ability to edit a draft stands apart from any vendor accountability for an error left in it.
Notes reach the chart by copy and paste. The clinician copies the whole note or a single section to fit the EHR layout, with PDF and Word export as a fallback when copying fails, and patient instructions and confidential adolescent content are handled separately. The help center tells clinicians to move the note before the 30 day deletion because the EHR is the record of care.
No connector to Epic, Oracle Health or any other EHR exists, and general blog posts about those systems are not integration documentation. The manual transfer is the limit, while the copy and export tools themselves are usable and fit different chart layouts.
Clinicians use the hosted service in a browser or the iOS and Android apps, with Google Cloud and MongoDB among the named infrastructure partners. A visit recorded without a connection stays encrypted on the device until upload, which is a temporary capture mode rather than local processing of transcription or generation.
No region commitment says where audio, transcripts, generated notes, backups and model processing stay. The Ashburn, Virginia address identifies the operator rather than a data boundary. Practice controls restrict who can see visits, but the full processing and support path across providers is not mapped, and the 30 day deletion window is the main limit on how long data sits anywhere.
Two live price lists disagree. The pricing page offers Individual at $149 a month with unlimited notes. The help center lists Individual at $40 a month for 40 notes and Unlimited at $100 a month or $1,080 a year, and its billing articles give practices of two to 25 clinicians a $100 monthly rate per clinician.
Both offer a 20 note trial with no card, larger groups get quotes, and students and trainees can ask for discounts. No notice says which paid schedule applies, so the trial is clear while the paid entry rate and the features attached to each plan are not.
Licensed clinicians and practices across North America are the intended users, with SOAP and progress formats for medical visits and DAP notes for behavioral health. In person visits, telehealth calls, uploaded recordings and dictation are all supported. Practice administration adds clinician roles and office staff access for referrals, and larger groups can run the product under their own brand.
The family medicine testimonial offers one named user's impression rather than a measured specialty evaluation, and educational posts about other fields do not validate the model there. Performance in noisy rooms, pediatric visits, specialty terminology and suggested differentials is not quantified.
Alternatives to SOAPsuds
Other products that do the same kind of work as SOAPsuds, and what each one does.
PatientNotes is an AI scribe for solo clinicians and small practices, co founded in 2023 by a dentist, Dr. Alisa, and her brother Vadim Kouznetsov, who runs the company from an address in London.
MedWriter is an AI medical scribe from S7Labs, LLC, doing business as Medwriter, built for outpatient clinicians and strongest in psychiatry, primary care, neurology and pain medicine.
Wavo Health is an AI scribe and clinical workspace sold to clinicians and independent practices in the United States and Canada by Wavo AI Technologies Inc., a Canadian federal corporation.
PMHScribe is an AI scribe built for psychiatric and behavioral health clinicians by Allison Sikorsky, a psychiatric mental health nurse practitioner and practice owner, and sold under its terms as PMHScribe by eRx.ai.
OneChart is an ambient AI scribe from a small Canadian seed stage team in Hamilton, Ontario. The scribe listens during the encounter, generates structured SOAP, DAP, BIRP and specialty notes, and returns them for clinician review and approval before export.
iOS first ambient scribe from OrbDoc, shipped as the OrbVoice app and aimed principally at solo clinicians and small practices, with an enterprise offering marketed alongside it.
Pricing
Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.
| Entry Price | Pricing Basis | BAA Tier | Implementation | Source |
|---|---|---|---|---|
|
20 note free trial; paid rates conflict
$0 baseline
|
The pricing page lists Individual at $149 per month for unlimited notes. The help center lists Individual at $40 per month for 40 notes and Unlimited at $100 per month or $1,080 per year. Neither schedule says it replaces the other. | A solo clinician or practice administrator accepts a BAA before the first visit. Contracted practices use the agreement in their contract, and no plan restricts it. | No separate implementation fee is listed. Configuration and white labeling for larger groups are quoted. | Vendor Published |
The free trial includes 20 notes without a payment card. The help center lists practices of two to 25 clinicians at $100 per clinician per month or $1,080 a year, and larger groups receive quotes. Students and trainees can ask for discounts. The pricing page and help center attach different note limits and prices to the Individual plan.